> ## Documentation Index
> Fetch the complete documentation index at: https://docs.sambanova.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# 対応モデルとバンドル

export const StackModelCard = ({provider = "", id = "", use = "", modalities = [], capabilities = [], endpoints = [], features = [], bundles = [], hf = "", customCheckpoints, draftModels = [], sequenceLengths = [], specDecoding = false}) => {
  const MARK_PATHS = {
    deepseek: `<path d="M23.748 4.651c-.254-.124-.364.113-.512.233-.051.04-.094.09-.137.137-.372.397-.806.657-1.373.626-.829-.046-1.537.214-2.163.848-.133-.782-.575-1.248-1.247-1.548-.352-.155-.708-.311-.955-.65-.172-.24-.219-.509-.305-.774-.055-.16-.11-.323-.293-.35-.2-.031-.278.136-.356.276-.313.572-.434 1.202-.422 1.84.027 1.436.633 2.58 1.838 3.393.137.094.172.187.129.323-.082.28-.18.553-.266.833-.055.179-.137.218-.328.14a5.5 5.5 0 0 1-1.737-1.179c-.857-.828-1.631-1.743-2.597-2.46a12 12 0 0 0-.689-.47c-.985-.957.13-1.743.387-1.836.27-.098.094-.433-.778-.428-.872.003-1.67.295-2.687.685a3 3 0 0 1-.465.136 9.6 9.6 0 0 0-2.883-.101c-1.885.21-3.39 1.1-4.497 2.622C.082 8.776-.231 10.854.152 13.02c.403 2.284 1.568 4.175 3.36 5.653 1.857 1.533 3.997 2.284 6.438 2.14 1.482-.085 3.132-.284 4.994-1.86.47.234.962.328 1.78.398.629.058 1.235-.031 1.705-.129.735-.155.684-.836.418-.961-2.155-1.004-1.682-.595-2.112-.926 1.095-1.295 2.768-3.598 3.284-6.733.05-.346.115-.834.108-1.114-.004-.171.035-.238.23-.257a4.2 4.2 0 0 0 1.545-.475c1.397-.763 1.96-2.016 2.093-3.517.02-.23-.004-.467-.247-.588M11.58 18.168c-2.088-1.642-3.101-2.183-3.52-2.16-.39.024-.32.472-.234.763.09.288.207.487.371.74.114.167.192.416-.113.603-.673.416-1.842-.14-1.897-.168-1.361-.801-2.5-1.86-3.301-3.306-.775-1.393-1.225-2.888-1.299-4.482-.02-.385.094-.522.477-.592a4.7 4.7 0 0 1 1.53-.038c2.131.311 3.946 1.264 5.467 2.774.868.86 1.525 1.887 2.202 2.89.72 1.066 1.494 2.082 2.48 2.915.348.291.626.513.892.677-.802.09-2.14.109-3.055-.615zm1.001-6.44a.306.306 0 0 1 .415-.287.3.3 0 0 1 .113.074.3.3 0 0 1 .086.214c0 .17-.136.307-.308.307a.303.303 0 0 1-.306-.307m3.11 1.596c-.2.081-.4.151-.591.16a1.25 1.25 0 0 1-.798-.254c-.274-.23-.47-.358-.551-.758a1.7 1.7 0 0 1 .015-.588c.07-.327-.007-.537-.238-.727-.188-.156-.426-.199-.689-.199a.6.6 0 0 1-.254-.078.253.253 0 0 1-.114-.358 1 1 0 0 1 .192-.21c.356-.202.767-.136 1.146.016.352.144.618.408 1.001.782.392.451.462.576.685.915.176.264.336.536.446.848.066.194-.02.353-.25.45"/>`,
    generic: `<path fill-rule="evenodd" clip-rule="evenodd" d="M12 1.5 21.5 7v10L12 22.5 2.5 17V7L12 1.5Zm0 2.31L4.5 8.16v7.68L12 20.19l7.5-4.35V8.16L12 3.81Z"/>`,
    google: `<path d="M12.48 10.92v3.28h7.84c-.24 1.84-.853 3.187-1.787 4.133-1.147 1.147-2.933 2.4-6.053 2.4-4.827 0-8.6-3.893-8.6-8.72s3.773-8.72 8.6-8.72c2.6 0 4.507 1.027 5.907 2.347l2.307-2.307C18.747 1.44 16.133 0 12.48 0 5.867 0 .307 5.387.307 12s5.56 12 12.173 12c3.573 0 6.267-1.173 8.373-3.36 2.16-2.16 2.84-5.213 2.84-7.667 0-.76-.053-1.467-.173-2.053H12.48z"/>`,
    meta: `<path d="M6.915 4.03c-1.968 0-3.683 1.28-4.871 3.113C.704 9.208 0 11.883 0 14.449c0 .706.07 1.369.21 1.973a6.624 6.624 0 0 0 .265.86 5.297 5.297 0 0 0 .371.761c.696 1.159 1.818 1.927 3.593 1.927 1.497 0 2.633-.671 3.965-2.444.76-1.012 1.144-1.626 2.663-4.32l.756-1.339.186-.325c.061.1.121.196.183.3l2.152 3.595c.724 1.21 1.665 2.556 2.47 3.314 1.046.987 1.992 1.22 3.06 1.22 1.075 0 1.876-.355 2.455-.843a3.743 3.743 0 0 0 .81-.973c.542-.939.861-2.127.861-3.745 0-2.72-.681-5.357-2.084-7.45-1.282-1.912-2.957-2.93-4.716-2.93-1.047 0-2.088.467-3.053 1.308-.652.57-1.257 1.29-1.82 2.05-.69-.875-1.335-1.547-1.958-2.056-1.182-.966-2.315-1.303-3.454-1.303zm10.16 2.053c1.147 0 2.188.758 2.992 1.999 1.132 1.748 1.647 4.195 1.647 6.4 0 1.548-.368 2.9-1.839 2.9-.58 0-1.027-.23-1.664-1.004-.496-.601-1.343-1.878-2.832-4.358l-.617-1.028a44.908 44.908 0 0 0-1.255-1.98c.07-.109.141-.224.211-.327 1.12-1.667 2.118-2.602 3.358-2.602zm-10.201.553c1.265 0 2.058.791 2.675 1.446.307.327.737.871 1.234 1.579l-1.02 1.566c-.757 1.163-1.882 3.017-2.837 4.338-1.191 1.649-1.81 1.817-2.486 1.817-.524 0-1.038-.237-1.383-.794-.263-.426-.464-1.13-.464-2.046 0-2.221.63-4.535 1.66-6.088.454-.687.964-1.226 1.533-1.533a2.264 2.264 0 0 1 1.088-.285z"/>`,
    minimax: `<path d="M11.43 3.92a.86.86 0 1 0-1.718 0v14.236a1.999 1.999 0 0 1-3.997 0V9.022a.86.86 0 1 0-1.718 0v3.87a1.999 1.999 0 0 1-3.997 0V11.49a.57.57 0 0 1 1.139 0v1.404a.86.86 0 0 0 1.719 0V9.022a1.999 1.999 0 0 1 3.997 0v9.134a.86.86 0 0 0 1.719 0V3.92a1.998 1.998 0 1 1 3.996 0v11.788a.57.57 0 1 1-1.139 0zm10.572 3.105a2 2 0 0 0-1.999 1.997v7.63a.86.86 0 0 1-1.718 0V3.923a1.999 1.999 0 0 0-3.997 0v16.16a.86.86 0 0 1-1.719 0V18.08a.57.57 0 1 0-1.138 0v2a1.998 1.998 0 0 0 3.996 0V3.92a.86.86 0 0 1 1.719 0v12.73a1.999 1.999 0 0 0 3.996 0V9.023a.86.86 0 1 1 1.72 0v6.686a.57.57 0 0 0 1.138 0V9.022a2 2 0 0 0-1.998-1.997"/>`,
    mistralai: `<path d="M17.143 3.429v3.428h-3.429v3.429h-3.428V6.857H6.857V3.43H3.43v13.714H0v3.428h10.286v-3.428H6.857v-3.429h3.429v3.429h3.429v-3.429h3.428v3.429h-3.428v3.428H24v-3.428h-3.43V3.429z"/>`,
    openai: `<path d="M22.2819 9.8211a5.9847 5.9847 0 0 0-.5157-4.9108 6.0462 6.0462 0 0 0-6.5098-2.9A6.0651 6.0651 0 0 0 4.9807 4.1818a5.9847 5.9847 0 0 0-3.9977 2.9 6.0462 6.0462 0 0 0 .7427 7.0966 5.98 5.98 0 0 0 .511 4.9107 6.051 6.051 0 0 0 6.5146 2.9001A5.9847 5.9847 0 0 0 13.2599 24a6.0557 6.0557 0 0 0 5.7718-4.2058 5.9894 5.9894 0 0 0 3.9977-2.9001 6.0557 6.0557 0 0 0-.7475-7.0729zm-9.022 12.6081a4.4755 4.4755 0 0 1-2.8764-1.0408l.1419-.0804 4.7783-2.7582a.7948.7948 0 0 0 .3927-.6813v-6.7369l2.02 1.1686a.071.071 0 0 1 .038.052v5.5826a4.504 4.504 0 0 1-4.4945 4.4944zm-9.6607-4.1254a4.4708 4.4708 0 0 1-.5346-3.0137l.142.0852 4.783 2.7582a.7712.7712 0 0 0 .7806 0l5.8428-3.3685v2.3324a.0804.0804 0 0 1-.0332.0615L9.74 19.9502a4.4992 4.4992 0 0 1-6.1408-1.6464zM2.3408 7.8956a4.485 4.485 0 0 1 2.3655-1.9728V11.6a.7664.7664 0 0 0 .3879.6765l5.8144 3.3543-2.0201 1.1685a.0757.0757 0 0 1-.071 0l-4.8303-2.7865A4.504 4.504 0 0 1 2.3408 7.872zm16.5963 3.8558L13.1038 8.364 15.1192 7.2a.0757.0757 0 0 1 .071 0l4.8303 2.7913a4.4944 4.4944 0 0 1-.6765 8.1042v-5.6772a.79.79 0 0 0-.407-.667zm2.0107-3.0231l-.142-.0852-4.7735-2.7818a.7759.7759 0 0 0-.7854 0L9.409 9.2297V6.8974a.0662.0662 0 0 1 .0284-.0615l4.8303-2.7866a4.4992 4.4992 0 0 1 6.6802 4.66zM8.3065 12.863l-2.02-1.1638a.0804.0804 0 0 1-.038-.0567V6.0742a4.4992 4.4992 0 0 1 7.3757-3.4537l-.142.0805L8.704 5.459a.7948.7948 0 0 0-.3927.6813zm1.0976-2.3654l2.602-1.4998 2.6069 1.4998v2.9994l-2.5974 1.4997-2.6067-1.4997Z"/>`,
    qwen: `<path d="M23.919 14.545 20.817 9.17l1.47-2.544a.56.56 0 0 0 0-.566l-1.633-2.83a.57.57 0 0 0-.49-.283h-6.207L12.487.402a.57.57 0 0 0-.49-.284H8.732a.56.56 0 0 0-.49.284L5.139 5.775h-2.94a.56.56 0 0 0-.49.284L.077 8.887a.56.56 0 0 0 0 .567L3.18 14.83l-1.47 2.545a.56.56 0 0 0 0 .566l1.634 2.83a.57.57 0 0 0 .49.283h6.205l1.47 2.545a.57.57 0 0 0 .49.284h3.266a.57.57 0 0 0 .49-.284l3.104-5.375h2.94a.57.57 0 0 0 .49-.283l1.634-2.828a.55.55 0 0 0-.004-.568M8.733.686l1.634 2.828-1.634 2.828H21.8L20.164 9.17H7.425L5.63 6.06Zm1.306 19.801-6.205-.002 1.634-2.83h3.265L2.201 6.344h3.267q3.182 5.517 6.367 11.032zm10.124-5.66L18.53 12l-6.532 11.315-1.634-2.83c2.129-3.673 4.25-7.351 6.373-11.028h3.592l3.102 5.374z"/>`
  };
  const PROVIDER_MARKS = {
    Meta: {
      icon: "meta",
      c: "#0467DF",
      d: "#2279E3"
    },
    MiniMax: {
      icon: "minimax",
      c: "#E73562",
      d: "#E73562"
    },
    "Mistral AI": {
      icon: "mistralai",
      c: "#FA520F",
      d: "#FA520F"
    },
    DeepSeek: {
      icon: "deepseek",
      c: "#5786FE",
      d: "#5786FE"
    },
    OpenAI: {
      icon: "openai",
      c: "#0D0D0D",
      d: "#F5F5F5"
    },
    Google: {
      icon: "google",
      c: "#4285F4",
      d: "#4285F4"
    },
    "Alibaba Cloud": {
      icon: "qwen",
      c: "#6950EF",
      d: "#7B65F1"
    }
  };
  const NEUTRAL = {
    icon: "generic",
    c: "#645D70",
    d: "#A9A1B6"
  };
  const mk = PROVIDER_MARKS[provider] || NEUTRAL;
  const dotFor = v => v === "Supported" ? "sst-dot sst-dot-ok" : v === "Limited" ? "sst-dot sst-dot-lim" : "sst-dot sst-dot-no";
  const featClass = v => v === "Supported" ? "sst-feat sst-feat-ok" : v === "Limited" ? "sst-feat sst-feat-lim" : "sst-feat";
  const slug = String(id).toLowerCase().replace(/[^a-z0-9_]+/g, "-").replace(/^-|-$/g, "");
  const [open, setOpen] = useState(false);
  useEffect(() => {
    const sync = () => {
      if (decodeURIComponent(window.location.hash).slice(1) === slug) setOpen(true);
    };
    sync();
    window.addEventListener("hashchange", sync);
    return () => window.removeEventListener("hashchange", sync);
  }, [slug]);
  return <div className="sst">
      <details className="sst-card sst-det" open={open} onToggle={e => setOpen(e.currentTarget.open)}>
        <summary className="sst-head">
          <div className="sst-head-l">
            <span className="sst-mark" style={{
    "--mk": mk.c,
    "--mkd": mk.d
  }} aria-hidden="true">
              <svg className="sst-mark-img" viewBox="0 0 24 24" fill="currentColor" role="img" aria-hidden="true" dangerouslySetInnerHTML={{
    __html: MARK_PATHS[mk.icon]
  }} />
            </span>
            <div className="sst-head-txt">
              <div className="sst-prov">{provider}</div>
              <div className="sst-id">{id}</div>
              {use && <p className="sst-use">{use}</p>}
            </div>
          </div>
          <svg className="sst-chev" viewBox="0 0 24 24" fill="none" stroke="currentColor" strokeWidth="2" strokeLinecap="round" strokeLinejoin="round" aria-hidden="true">
            <path d="m6 9 6 6 6-6" />
          </svg>
        </summary>

        <div className="sst-body">
          <div>
            <p className="sst-lbl">Modalities</p>
            <div className="sst-row">
              {modalities.length ? modalities.map(x => <span className="sst-tag sst-tag-mod" key={x}>
                    {x}
                  </span>) : <span className="sst-tag">—</span>}
            </div>
          </div>

          {capabilities.length > 0 && <div>
              <p className="sst-lbl">Capabilities</p>
              <div className="sst-row">
                {capabilities.map(x => <span className="sst-tag" key={x}>
                    {x}
                  </span>)}
              </div>
            </div>}

          {endpoints.length > 0 && <div>
              <p className="sst-lbl">Endpoints</p>
              <div className="sst-eps">
                {endpoints.map(([name, path, verdict]) => <div className={verdict ? "sst-ep" : "sst-ep sst-ep-off"} key={name}>
                    <span className={dotFor(verdict)} />
                    <div className="sst-ep-t">
                      <div className="sst-ep-n">{name}</div>
                      <div className="sst-ep-p">{path}</div>
                    </div>
                  </div>)}
              </div>
            </div>}

          {features.length > 0 && <div>
              <p className="sst-lbl">Features</p>
              <div className="sst-feats">
                {features.map(([name, verdict]) => <span className={featClass(verdict)} key={name}>
                    <span className={dotFor(verdict)} />
                    {name}
                    {verdict === "Limited" && " (limited)"}
                  </span>)}
              </div>
            </div>}

          {bundles.length > 0 && <div>
              <p className="sst-lbl">Bundle configurations</p>
              <div className="sst-bundles">
                {bundles.map(([name, ctx, isSuggested, status]) => <div className={isSuggested ? "sst-bundle sst-bundle-sug" : "sst-bundle"} key={name}>
                    <span className="sst-bn">{name}</span>
                    <span className="sst-stat">{ctx}</span>
                    {status && <span className={"sst-pef sst-pef-" + status} title={"PEF status: " + status}>
                        {status}
                      </span>}
                    {isSuggested && <span className="sst-sug">Suggested</span>}
                  </div>)}
              </div>
            </div>}

          {sequenceLengths.length > 0 && <div>
              <p className="sst-lbl">Context length (batch size)</p>
              <div className="sst-bundles">
                {sequenceLengths.map(sl => <div className="sst-bundle" key={sl}>
                    <span className="sst-bn">{sl}</span>
                  </div>)}
              </div>
            </div>}

          {(typeof customCheckpoints === "boolean" || draftModels.length > 0 || specDecoding) && <div>
              <p className="sst-lbl">Deployment options</p>
              <div className="sst-opts">
                {typeof customCheckpoints === "boolean" && <div className="sst-opt">
                    <span className="sst-opt-k">Custom checkpoints (BYOC)</span>
                    <span className="sst-opt-v">{customCheckpoints ? "Supported" : "Not supported"}</span>
                  </div>}
                {draftModels.length > 0 ? <div className="sst-opt">
                    <span className="sst-opt-k">Speculative decoding</span>
                    <span className="sst-opt-v">
                      Draft {draftModels.length > 1 ? "models" : "model"}:{" "}
                      {draftModels.map(d => <code key={d}>{d}</code>)}
                    </span>
                  </div> : specDecoding ? <div className="sst-opt">
                    <span className="sst-opt-k">Speculative decoding</span>
                    <span className="sst-opt-v">Supported</span>
                  </div> : null}
              </div>
            </div>}

          {hf && <a className="sst-link" href={hf} target="_blank" rel="noreferrer">
              View model card on Hugging Face ↗
            </a>}
        </div>
      </details>
    </div>;
};

export const StackModelExplorer = ({models = []}) => {
  const BRAND = "#974fc7";
  const MARK_PATHS = {
    deepseek: `<path d="M23.748 4.651c-.254-.124-.364.113-.512.233-.051.04-.094.09-.137.137-.372.397-.806.657-1.373.626-.829-.046-1.537.214-2.163.848-.133-.782-.575-1.248-1.247-1.548-.352-.155-.708-.311-.955-.65-.172-.24-.219-.509-.305-.774-.055-.16-.11-.323-.293-.35-.2-.031-.278.136-.356.276-.313.572-.434 1.202-.422 1.84.027 1.436.633 2.58 1.838 3.393.137.094.172.187.129.323-.082.28-.18.553-.266.833-.055.179-.137.218-.328.14a5.5 5.5 0 0 1-1.737-1.179c-.857-.828-1.631-1.743-2.597-2.46a12 12 0 0 0-.689-.47c-.985-.957.13-1.743.387-1.836.27-.098.094-.433-.778-.428-.872.003-1.67.295-2.687.685a3 3 0 0 1-.465.136 9.6 9.6 0 0 0-2.883-.101c-1.885.21-3.39 1.1-4.497 2.622C.082 8.776-.231 10.854.152 13.02c.403 2.284 1.568 4.175 3.36 5.653 1.857 1.533 3.997 2.284 6.438 2.14 1.482-.085 3.132-.284 4.994-1.86.47.234.962.328 1.78.398.629.058 1.235-.031 1.705-.129.735-.155.684-.836.418-.961-2.155-1.004-1.682-.595-2.112-.926 1.095-1.295 2.768-3.598 3.284-6.733.05-.346.115-.834.108-1.114-.004-.171.035-.238.23-.257a4.2 4.2 0 0 0 1.545-.475c1.397-.763 1.96-2.016 2.093-3.517.02-.23-.004-.467-.247-.588M11.58 18.168c-2.088-1.642-3.101-2.183-3.52-2.16-.39.024-.32.472-.234.763.09.288.207.487.371.74.114.167.192.416-.113.603-.673.416-1.842-.14-1.897-.168-1.361-.801-2.5-1.86-3.301-3.306-.775-1.393-1.225-2.888-1.299-4.482-.02-.385.094-.522.477-.592a4.7 4.7 0 0 1 1.53-.038c2.131.311 3.946 1.264 5.467 2.774.868.86 1.525 1.887 2.202 2.89.72 1.066 1.494 2.082 2.48 2.915.348.291.626.513.892.677-.802.09-2.14.109-3.055-.615zm1.001-6.44a.306.306 0 0 1 .415-.287.3.3 0 0 1 .113.074.3.3 0 0 1 .086.214c0 .17-.136.307-.308.307a.303.303 0 0 1-.306-.307m3.11 1.596c-.2.081-.4.151-.591.16a1.25 1.25 0 0 1-.798-.254c-.274-.23-.47-.358-.551-.758a1.7 1.7 0 0 1 .015-.588c.07-.327-.007-.537-.238-.727-.188-.156-.426-.199-.689-.199a.6.6 0 0 1-.254-.078.253.253 0 0 1-.114-.358 1 1 0 0 1 .192-.21c.356-.202.767-.136 1.146.016.352.144.618.408 1.001.782.392.451.462.576.685.915.176.264.336.536.446.848.066.194-.02.353-.25.45"/>`,
    generic: `<path fill-rule="evenodd" clip-rule="evenodd" d="M12 1.5 21.5 7v10L12 22.5 2.5 17V7L12 1.5Zm0 2.31L4.5 8.16v7.68L12 20.19l7.5-4.35V8.16L12 3.81Z"/>`,
    google: `<path d="M12.48 10.92v3.28h7.84c-.24 1.84-.853 3.187-1.787 4.133-1.147 1.147-2.933 2.4-6.053 2.4-4.827 0-8.6-3.893-8.6-8.72s3.773-8.72 8.6-8.72c2.6 0 4.507 1.027 5.907 2.347l2.307-2.307C18.747 1.44 16.133 0 12.48 0 5.867 0 .307 5.387.307 12s5.56 12 12.173 12c3.573 0 6.267-1.173 8.373-3.36 2.16-2.16 2.84-5.213 2.84-7.667 0-.76-.053-1.467-.173-2.053H12.48z"/>`,
    meta: `<path d="M6.915 4.03c-1.968 0-3.683 1.28-4.871 3.113C.704 9.208 0 11.883 0 14.449c0 .706.07 1.369.21 1.973a6.624 6.624 0 0 0 .265.86 5.297 5.297 0 0 0 .371.761c.696 1.159 1.818 1.927 3.593 1.927 1.497 0 2.633-.671 3.965-2.444.76-1.012 1.144-1.626 2.663-4.32l.756-1.339.186-.325c.061.1.121.196.183.3l2.152 3.595c.724 1.21 1.665 2.556 2.47 3.314 1.046.987 1.992 1.22 3.06 1.22 1.075 0 1.876-.355 2.455-.843a3.743 3.743 0 0 0 .81-.973c.542-.939.861-2.127.861-3.745 0-2.72-.681-5.357-2.084-7.45-1.282-1.912-2.957-2.93-4.716-2.93-1.047 0-2.088.467-3.053 1.308-.652.57-1.257 1.29-1.82 2.05-.69-.875-1.335-1.547-1.958-2.056-1.182-.966-2.315-1.303-3.454-1.303zm10.16 2.053c1.147 0 2.188.758 2.992 1.999 1.132 1.748 1.647 4.195 1.647 6.4 0 1.548-.368 2.9-1.839 2.9-.58 0-1.027-.23-1.664-1.004-.496-.601-1.343-1.878-2.832-4.358l-.617-1.028a44.908 44.908 0 0 0-1.255-1.98c.07-.109.141-.224.211-.327 1.12-1.667 2.118-2.602 3.358-2.602zm-10.201.553c1.265 0 2.058.791 2.675 1.446.307.327.737.871 1.234 1.579l-1.02 1.566c-.757 1.163-1.882 3.017-2.837 4.338-1.191 1.649-1.81 1.817-2.486 1.817-.524 0-1.038-.237-1.383-.794-.263-.426-.464-1.13-.464-2.046 0-2.221.63-4.535 1.66-6.088.454-.687.964-1.226 1.533-1.533a2.264 2.264 0 0 1 1.088-.285z"/>`,
    minimax: `<path d="M11.43 3.92a.86.86 0 1 0-1.718 0v14.236a1.999 1.999 0 0 1-3.997 0V9.022a.86.86 0 1 0-1.718 0v3.87a1.999 1.999 0 0 1-3.997 0V11.49a.57.57 0 0 1 1.139 0v1.404a.86.86 0 0 0 1.719 0V9.022a1.999 1.999 0 0 1 3.997 0v9.134a.86.86 0 0 0 1.719 0V3.92a1.998 1.998 0 1 1 3.996 0v11.788a.57.57 0 1 1-1.139 0zm10.572 3.105a2 2 0 0 0-1.999 1.997v7.63a.86.86 0 0 1-1.718 0V3.923a1.999 1.999 0 0 0-3.997 0v16.16a.86.86 0 0 1-1.719 0V18.08a.57.57 0 1 0-1.138 0v2a1.998 1.998 0 0 0 3.996 0V3.92a.86.86 0 0 1 1.719 0v12.73a1.999 1.999 0 0 0 3.996 0V9.023a.86.86 0 1 1 1.72 0v6.686a.57.57 0 0 0 1.138 0V9.022a2 2 0 0 0-1.998-1.997"/>`,
    mistralai: `<path d="M17.143 3.429v3.428h-3.429v3.429h-3.428V6.857H6.857V3.43H3.43v13.714H0v3.428h10.286v-3.428H6.857v-3.429h3.429v3.429h3.429v-3.429h3.428v3.429h-3.428v3.428H24v-3.428h-3.43V3.429z"/>`,
    openai: `<path d="M22.2819 9.8211a5.9847 5.9847 0 0 0-.5157-4.9108 6.0462 6.0462 0 0 0-6.5098-2.9A6.0651 6.0651 0 0 0 4.9807 4.1818a5.9847 5.9847 0 0 0-3.9977 2.9 6.0462 6.0462 0 0 0 .7427 7.0966 5.98 5.98 0 0 0 .511 4.9107 6.051 6.051 0 0 0 6.5146 2.9001A5.9847 5.9847 0 0 0 13.2599 24a6.0557 6.0557 0 0 0 5.7718-4.2058 5.9894 5.9894 0 0 0 3.9977-2.9001 6.0557 6.0557 0 0 0-.7475-7.0729zm-9.022 12.6081a4.4755 4.4755 0 0 1-2.8764-1.0408l.1419-.0804 4.7783-2.7582a.7948.7948 0 0 0 .3927-.6813v-6.7369l2.02 1.1686a.071.071 0 0 1 .038.052v5.5826a4.504 4.504 0 0 1-4.4945 4.4944zm-9.6607-4.1254a4.4708 4.4708 0 0 1-.5346-3.0137l.142.0852 4.783 2.7582a.7712.7712 0 0 0 .7806 0l5.8428-3.3685v2.3324a.0804.0804 0 0 1-.0332.0615L9.74 19.9502a4.4992 4.4992 0 0 1-6.1408-1.6464zM2.3408 7.8956a4.485 4.485 0 0 1 2.3655-1.9728V11.6a.7664.7664 0 0 0 .3879.6765l5.8144 3.3543-2.0201 1.1685a.0757.0757 0 0 1-.071 0l-4.8303-2.7865A4.504 4.504 0 0 1 2.3408 7.872zm16.5963 3.8558L13.1038 8.364 15.1192 7.2a.0757.0757 0 0 1 .071 0l4.8303 2.7913a4.4944 4.4944 0 0 1-.6765 8.1042v-5.6772a.79.79 0 0 0-.407-.667zm2.0107-3.0231l-.142-.0852-4.7735-2.7818a.7759.7759 0 0 0-.7854 0L9.409 9.2297V6.8974a.0662.0662 0 0 1 .0284-.0615l4.8303-2.7866a4.4992 4.4992 0 0 1 6.6802 4.66zM8.3065 12.863l-2.02-1.1638a.0804.0804 0 0 1-.038-.0567V6.0742a4.4992 4.4992 0 0 1 7.3757-3.4537l-.142.0805L8.704 5.459a.7948.7948 0 0 0-.3927.6813zm1.0976-2.3654l2.602-1.4998 2.6069 1.4998v2.9994l-2.5974 1.4997-2.6067-1.4997Z"/>`,
    qwen: `<path d="M23.919 14.545 20.817 9.17l1.47-2.544a.56.56 0 0 0 0-.566l-1.633-2.83a.57.57 0 0 0-.49-.283h-6.207L12.487.402a.57.57 0 0 0-.49-.284H8.732a.56.56 0 0 0-.49.284L5.139 5.775h-2.94a.56.56 0 0 0-.49.284L.077 8.887a.56.56 0 0 0 0 .567L3.18 14.83l-1.47 2.545a.56.56 0 0 0 0 .566l1.634 2.83a.57.57 0 0 0 .49.283h6.205l1.47 2.545a.57.57 0 0 0 .49.284h3.266a.57.57 0 0 0 .49-.284l3.104-5.375h2.94a.57.57 0 0 0 .49-.283l1.634-2.828a.55.55 0 0 0-.004-.568M8.733.686l1.634 2.828-1.634 2.828H21.8L20.164 9.17H7.425L5.63 6.06Zm1.306 19.801-6.205-.002 1.634-2.83h3.265L2.201 6.344h3.267q3.182 5.517 6.367 11.032zm10.124-5.66L18.53 12l-6.532 11.315-1.634-2.83c2.129-3.673 4.25-7.351 6.373-11.028h3.592l3.102 5.374z"/>`
  };
  const PROVIDER_MARKS = {
    Meta: {
      icon: "meta",
      c: "#0467DF",
      d: "#2279E3"
    },
    MiniMax: {
      icon: "minimax",
      c: "#E73562",
      d: "#E73562"
    },
    "Mistral AI": {
      icon: "mistralai",
      c: "#FA520F",
      d: "#FA520F"
    },
    DeepSeek: {
      icon: "deepseek",
      c: "#5786FE",
      d: "#5786FE"
    },
    OpenAI: {
      icon: "openai",
      c: "#0D0D0D",
      d: "#F5F5F5"
    },
    Google: {
      icon: "google",
      c: "#4285F4",
      d: "#4285F4"
    },
    "Alibaba Cloud": {
      icon: "qwen",
      c: "#6950EF",
      d: "#7B65F1"
    }
  };
  const NEUTRAL = {
    icon: "generic",
    c: "#645D70",
    d: "#A9A1B6"
  };
  const markFor = p => PROVIDER_MARKS[p] || NEUTRAL;
  const CSS = `
  .sst { --sst-bd: rgba(0,0,0,.10); --sst-mu: rgba(0,0,0,.55); --sst-sf: rgba(0,0,0,.025); --sst-sf2: rgba(0,0,0,.05); }
  .dark .sst { --sst-bd: rgba(255,255,255,.14); --sst-mu: rgba(255,255,255,.60); --sst-sf: rgba(255,255,255,.04); --sst-sf2: rgba(255,255,255,.07); }

  /* ---- controls ---- */
  .sst-controls { display:flex; flex-wrap:wrap; gap:10px; align-items:center; margin:4px 0 14px; }
  .sst-search { flex:1 1 300px; min-width:200px; font:inherit; font-size:14px; padding:10px 14px; border:1px solid var(--sst-bd); border-radius:10px; background:transparent; color:inherit; }
  .sst-search:focus { outline:none; border-color:${BRAND}; box-shadow:0 0 0 3px rgba(151,79,199,.14); }
  .sst-chips { display:flex; flex-wrap:wrap; gap:7px; margin-bottom:10px; }
  .sst-chip { display:inline-flex; align-items:center; gap:7px; font:inherit; font-size:12.5px; padding:5px 12px 5px 6px; border:1px solid var(--sst-bd); border-radius:999px; background:transparent; color:inherit; cursor:pointer; transition:.12s; }
  .sst-chip-plain { padding-left:12px; }
  .sst-chip:hover { border-color:${BRAND}; }
  .sst-chip-on { border-color:${BRAND}; background:rgba(151,79,199,.12); color:${BRAND}; font-weight:600; }
  .sst-chip-n { opacity:.55; font-variant-numeric:tabular-nums; }
  .sst-count { font-size:13px; color:var(--sst-mu); margin:0 0 14px; }

  /* ---- overview grid ---- */
  .sst-grid { display:grid; grid-template-columns:repeat(auto-fill,minmax(260px,1fr)); gap:12px; }
  .sst-mini { display:flex; flex-direction:column; gap:9px; padding:14px; border:1px solid var(--sst-bd); border-radius:12px; background:var(--sst-sf); text-decoration:none!important; color:inherit!important; transition:.12s; }
  .sst-mini:hover { border-color:${BRAND}; background:var(--sst-sf2); transform:translateY(-1px); }
  .sst-mini-top { display:flex; justify-content:space-between; align-items:center; gap:8px; }
  .sst-prov { font-size:10.5px; font-weight:700; letter-spacing:.06em; text-transform:uppercase; color:var(--sst-mu); }
  .sst-mini-id { font-family:ui-monospace,SFMono-Regular,Menlo,monospace; font-size:13px; font-weight:600; word-break:break-all; line-height:1.35; }
  .sst-mini-foot { display:flex; flex-wrap:wrap; gap:6px; align-items:center; margin-top:auto; }

  /* ---- provider mark ---- */
  .sst-mark {
    width:34px; height:34px; flex-shrink:0;
    display:inline-flex; align-items:center; justify-content:center;
    font-family:ui-monospace,SFMono-Regular,Menlo,monospace;
    font-size:12px; font-weight:700; letter-spacing:-.02em; line-height:1;
    border-radius:9px;
    color:var(--mk);
    background:color-mix(in srgb, var(--mk) 13%, transparent);
    border:1px solid color-mix(in srgb, var(--mk) 24%, transparent);
  }
  .dark .sst-mark {
    color:var(--mkd);
    background:color-mix(in srgb, var(--mkd) 15%, transparent);
    border-color:color-mix(in srgb, var(--mkd) 28%, transparent);
  }
  .sst-mark-sm { width:20px; height:20px; border-radius:6px; font-size:9px; }
  /* Inline brand marks. The SVG is emitted with fill="currentColor", so the
     glyph takes the colour set on the tile — which is the provider hue, with a
     separate value for dark mode. A mask was used before, against an external
     SVG; that broke under the production /docs base path. */
  .sst-mark-img { width:18px; height:18px; display:block; }
  .sst-mark-sm .sst-mark-img { width:11px; height:11px; }

  /* ---- badges ---- */

  .sst-stat { font-size:11.5px; color:var(--sst-mu); font-variant-numeric:tabular-nums; }
  .sst-tag { font-size:11px; padding:2px 8px; border-radius:999px; border:1px solid var(--sst-bd); color:var(--sst-mu); white-space:nowrap; }
  .sst-tag-mod { border-color:transparent; background:var(--sst-sf2); color:inherit; font-weight:600; }

  /* ---- detail card ---- */
  .sst-card { border:1px solid var(--sst-bd); border-radius:14px; overflow:hidden; margin:2px 0 8px; }
  .sst-head { display:flex; flex-wrap:wrap; justify-content:space-between; align-items:flex-start; gap:12px; padding:16px 18px; background:var(--sst-sf); border-bottom:1px solid var(--sst-bd); }
  .sst-head-l { min-width:0; display:flex; gap:12px; align-items:flex-start; }
  .sst-head-txt { min-width:0; }
  .sst-id { font-family:ui-monospace,SFMono-Regular,Menlo,monospace; font-size:15px; font-weight:700; word-break:break-all; line-height:1.3; }
  .sst-use { font-size:13px; color:var(--sst-mu); margin:5px 0 0; }

  /* Collapsible card.

     The full catalogue runs to several screens of cards, which pushed the
     filter controls far off the top of the page. Each card is a <details>:
     the header — provider, model ID, and use — always shows, and the detail
     below opens on click. Native <details> keeps this keyboard
     accessible and lets Chrome's find-in-page open a closed card. */
  .sst-det > summary { cursor:pointer; list-style:none; }
  .sst-det > summary::-webkit-details-marker { display:none; }
  .sst-det > summary:hover { background:var(--sst-sf2); }
  .sst-det:not([open]) > summary { border-bottom:none; }
  .sst-det > summary:focus-visible { outline:2px solid ${BRAND}; outline-offset:-2px; }
  .sst-chev { flex-shrink:0; width:18px; height:18px; margin-top:3px; color:var(--sst-mu); transition:transform .15s ease; }
  .sst-det[open] .sst-chev { transform:rotate(180deg); }

  .sst-body { padding:16px 18px; display:flex; flex-direction:column; gap:16px; }
  .sst-row { display:flex; flex-wrap:wrap; gap:7px; align-items:center; }
  .sst-lbl { font-size:10.5px; font-weight:700; letter-spacing:.07em; text-transform:uppercase; color:var(--sst-mu); margin:0 0 7px; }

  /* endpoints */
  .sst-eps { display:grid; grid-template-columns:repeat(auto-fill,minmax(215px,1fr)); gap:8px; }
  .sst-ep { display:flex; align-items:center; gap:9px; padding:9px 11px; border:1px solid var(--sst-bd); border-radius:9px; background:var(--sst-sf); }
  .sst-ep-off { opacity:.5; }
  .sst-dot { width:7px; height:7px; border-radius:50%; flex-shrink:0; }
  .sst-dot-ok { background:#059669; }
  .sst-dot-lim { background:#d97706; }
  .sst-dot-no { background:currentColor; opacity:.3; }
  .sst-ep-t { min-width:0; }
  .sst-ep-n { font-size:12.5px; font-weight:600; line-height:1.3; }
  .sst-ep-p { font-family:ui-monospace,SFMono-Regular,Menlo,monospace; font-size:10.5px; color:var(--sst-mu); word-break:break-all; }

  /* features */
  .sst-feats { display:flex; flex-wrap:wrap; gap:7px; }
  .sst-feat { display:inline-flex; align-items:center; gap:6px; font-size:12px; padding:5px 10px; border-radius:999px; border:1px solid var(--sst-bd); }
  .sst-feat-ok { border-color:rgba(5,150,105,.4); background:rgba(5,150,105,.08); }
  .sst-feat-lim { border-color:rgba(217,119,6,.4); background:rgba(217,119,6,.08); }

  /* bundles */
  .sst-bundles { display:flex; flex-direction:column; gap:7px; }
  .sst-bundle { display:flex; flex-wrap:wrap; align-items:center; gap:10px; padding:10px 12px; border:1px solid var(--sst-bd); border-radius:9px; }
  .sst-bundle-sug { border-color:${BRAND}; background:rgba(151,79,199,.06); }
  .sst-bn { font-family:ui-monospace,SFMono-Regular,Menlo,monospace; font-size:12px; font-weight:600; word-break:break-all; flex:1 1 190px; }
  .sst-sug { font-size:10px; font-weight:700; letter-spacing:.04em; text-transform:uppercase; color:${BRAND}; background:rgba(151,79,199,.13); padding:3px 7px; border-radius:5px; white-space:nowrap; }

  /* PEF status, from the release's Pef custom resources. Colour-independent
     readers still get the word, so the hue is reinforcement rather than the
     only signal. Alpha backgrounds keep these legible on both themes. */
  .sst-pef { font-size:10px; font-weight:700; letter-spacing:.04em; text-transform:uppercase; padding:3px 7px; border-radius:5px; white-space:nowrap; }
  .sst-pef-stable { color:#15803d; background:rgba(21,128,61,.13); }
  .sst-pef-preview { color:#b45309; background:rgba(180,83,9,.14); }
  .sst-pef-deprecated { color:#b91c1c; background:rgba(185,28,28,.13); }
  :root[data-theme="dark"] .sst-pef-stable, html.dark .sst-pef-stable { color:#4ade80; background:rgba(74,222,128,.15); }
  :root[data-theme="dark"] .sst-pef-preview, html.dark .sst-pef-preview { color:#fbbf24; background:rgba(251,191,36,.16); }
  :root[data-theme="dark"] .sst-pef-deprecated, html.dark .sst-pef-deprecated { color:#f87171; background:rgba(248,113,113,.15); }

  /* deployment options — BYOC and speculative decoding */
  .sst-opts { display:flex; flex-direction:column; gap:6px; }
  .sst-opt { display:flex; flex-wrap:wrap; align-items:baseline; gap:8px; font-size:13px; }
  .sst-opt-k { color:var(--sst-mu); flex:0 0 auto; }
  .sst-opt-v { font-weight:600; }
  .sst-opt-v code { font-family:ui-monospace,SFMono-Regular,Menlo,monospace; font-size:12px; font-weight:600; margin-right:6px; }

  .sst-link { font-size:13px; }
  .sst-empty { font-size:14px; color:var(--sst-mu); padding:20px 0; }

  @media (max-width:640px) {
    .sst-grid { grid-template-columns:1fr; }
    .sst-eps { grid-template-columns:1fr; }
  }
  `;
  const [q, setQ] = useState("");
  const [prov, setProv] = useState("all");
  const [mode, setMode] = useState("all");
  const [feat, setFeat] = useState("all");
  const FEATURES = [{
    key: "byoc",
    label: "Custom checkpoints"
  }, {
    key: "specdec",
    label: "Speculative decoding"
  }];
  const needle = q.trim().toLowerCase();
  const haystack = m => [m.id, m.provider, m.use, (m.modalities || []).join(" "), (m.capabilities || []).join(" "), m.byoc ? "byoc custom checkpoints" : "", m.specdec ? "speculative decoding draft model" : ""].join(" ").toLowerCase();
  const matchesFeat = m => feat === "all" || m[feat] === true;
  const afterSearch = models.filter(m => !needle || haystack(m).includes(needle));
  const byProv = afterSearch.filter(m => (mode === "all" || (m.modalities || []).includes(mode)) && matchesFeat(m));
  const byMode = afterSearch.filter(m => (prov === "all" || m.provider === prov) && matchesFeat(m));
  const byFeat = afterSearch.filter(m => (prov === "all" || m.provider === prov) && (mode === "all" || (m.modalities || []).includes(mode)));
  const shown = afterSearch.filter(m => (prov === "all" || m.provider === prov) && (mode === "all" || (m.modalities || []).includes(mode)) && matchesFeat(m));
  const providers = [...new Set(models.map(m => m.provider))].sort();
  const MODALITY_ORDER = ["Text", "Image", "Video", "Audio", "Embedding"];
  const present = new Set(models.flatMap(m => m.modalities || []));
  const modalities = [...MODALITY_ORDER.filter(x => present.has(x)), ...[...present].filter(x => !MODALITY_ORDER.includes(x)).sort()];
  const anchor = s => String(s).toLowerCase().replace(/[^a-z0-9_]+/g, "-").replace(/^-|-$/g, "");
  return <div className="sst">
      <style dangerouslySetInnerHTML={{
    __html: CSS
  }} />

      <div className="sst-controls">
        <input type="text" className="sst-search" value={q} onChange={e => setQ(e.target.value)} placeholder="Search by model ID, provider, or capability…" aria-label="Search models" />
      </div>

      <div className="sst-chips" role="group" aria-label="Filter by provider">
        <button type="button" className={prov === "all" ? "sst-chip sst-chip-plain sst-chip-on" : "sst-chip sst-chip-plain"} onClick={() => setProv("all")} aria-pressed={prov === "all"}>
          All providers <span className="sst-chip-n">{byProv.length}</span>
        </button>
        {providers.map(pName => {
    const n = byProv.filter(m => m.provider === pName).length;
    return <button key={pName} type="button" className={prov === pName ? "sst-chip sst-chip-on" : "sst-chip"} onClick={() => setProv(pName)} aria-pressed={prov === pName}>
              <span className="sst-mark sst-mark-sm" style={{
      "--mk": markFor(pName).c,
      "--mkd": markFor(pName).d
    }} aria-hidden="true">
                <svg className="sst-mark-img" viewBox="0 0 24 24" fill="currentColor" role="img" aria-hidden="true" dangerouslySetInnerHTML={{
      __html: MARK_PATHS[markFor(pName).icon]
    }} />
              </span>
              {pName} <span className="sst-chip-n">{n}</span>
            </button>;
  })}
      </div>

      <div className="sst-chips" role="group" aria-label="Filter by modality">
        <button type="button" className={mode === "all" ? "sst-chip sst-chip-plain sst-chip-on" : "sst-chip sst-chip-plain"} onClick={() => setMode("all")} aria-pressed={mode === "all"}>
          All modalities <span className="sst-chip-n">{byMode.length}</span>
        </button>
        {modalities.map(x => {
    const n = byMode.filter(m => (m.modalities || []).includes(x)).length;
    return <button key={x} type="button" className={mode === x ? "sst-chip sst-chip-plain sst-chip-on" : "sst-chip sst-chip-plain"} onClick={() => setMode(x)} aria-pressed={mode === x}>
              {x} <span className="sst-chip-n">{n}</span>
            </button>;
  })}
      </div>

      {}
      {FEATURES.some(f => models.some(m => m[f.key] === true)) && <div className="sst-chips" role="group" aria-label="Filter by deployment option">
          <button type="button" className={feat === "all" ? "sst-chip sst-chip-plain sst-chip-on" : "sst-chip sst-chip-plain"} onClick={() => setFeat("all")} aria-pressed={feat === "all"}>
            All options <span className="sst-chip-n">{byFeat.length}</span>
          </button>
          {FEATURES.filter(f => models.some(m => m[f.key] === true)).map(f => {
    const n = byFeat.filter(m => m[f.key] === true).length;
    return <button key={f.key} type="button" className={feat === f.key ? "sst-chip sst-chip-plain sst-chip-on" : "sst-chip sst-chip-plain"} onClick={() => setFeat(f.key)} aria-pressed={feat === f.key}>
                {f.label} <span className="sst-chip-n">{n}</span>
              </button>;
  })}
        </div>}

      <p className="sst-count">
        {shown.length} {shown.length === 1 ? "model" : "models"} — select one to jump to its full
        configuration below
      </p>

      <div className="sst-grid">
        {shown.map(m => <a className="sst-mini" key={m.id} href={`#${anchor(m.id)}`}>
            <div className="sst-mini-top">
              <span className="sst-mark" style={{
    "--mk": markFor(m.provider).c,
    "--mkd": markFor(m.provider).d
  }} aria-hidden="true">
                <svg className="sst-mark-img" viewBox="0 0 24 24" fill="currentColor" role="img" aria-hidden="true" dangerouslySetInnerHTML={{
    __html: MARK_PATHS[markFor(m.provider).icon]
  }} />
              </span>
            </div>
            <span className="sst-prov">{m.provider}</span>
            <span className="sst-mini-id">{m.id}</span>
            <div className="sst-row">
              {(m.modalities || []).map(x => <span className="sst-tag sst-tag-mod" key={x}>
                  {x}
                </span>)}
            </div>
            <div className="sst-mini-foot">
              {m.context && m.context !== "\u2014" && <span className="sst-stat">{m.context} (max context length)</span>}
            </div>
          </a>)}
      </div>

      {shown.length === 0 && <p className="sst-empty">No models match those filters. Clear the search, or reset a filter above.</p>}
    </div>;
};

SambaStack は、オンプレミス環境とホステッド環境の両方にデプロイできるさまざまなモデルをサポートしています。お使いのデプロイメントでどのモデルが利用可能かは、システム管理者にお問い合わせください。**Preview** (プレビュー) および **Stable** (安定版) の区分の定義、モデルがステージ間を移行する流れ、モデルの提供終了 (Retired) 前に届く通知、および提供終了の履歴については、[SambaStack モデルライフサイクル](/docs/ja/sambastack/service-administration/model-deployment/model-lifecycle)をご参照ください。

<Note>
  **カスタムチェックポイント**は、decoder-only のテキスト生成モデルすべてでサポートされています。ただし、以下の例外があります。

  | モデル | 理由 |
  | - | - |
  | `Llama-4-Maverick-17B-128E-Instruct`, `MiniMax-M3`, `gemma-3-27b-it`, `gemma-3-12b-it`, `gemma-4-31B-it` | マルチモーダルであり、decoder-only のテキスト生成モデルではない |
  | `Qwen3-32B` | 変換パッケージが未提供 |
  | `Whisper-Large-v3`, `Qwen3-TTS-Talker`, `Qwen3-TTS-Vocoder`, `E5-Mistral-7B-Instruct` | テキスト生成モデルではない |

  チェックポイントの変換方法は、[チェックポイント変換ツール](/docs/ja/sambastack/service-administration/model-deployment/checkpoint-conversion-tool)をご参照ください。
</Note>

## デプロイオプション

SambaStack でモデルをデプロイする際、管理者はさまざまなコンテキスト長とバッチサイズの組み合わせから選択できます。

* バッチサイズが小さいほど、トークンスループット (トークン/秒) が高くなります。
* バッチサイズが大きいほど、複数ユーザーに対する同時実行性が向上します。

SambaStack は、高インタラクティビティ (high-interactivity) と高スループット (high-throughput) の 2 種類のデプロイ構成をサポートしています。それぞれの使い分けについては、[高スループットデプロイ](/docs/ja/sambastack/service-administration/model-deployment/deploying-models-and-bundles/deployment-configurations)をご参照ください。

## クラスター上のモデルの確認

クラスターで利用可能なモデルは、次のコマンドで確認できます。

```shellscript theme={}
kubectl -n <namespace> get models
```

<Accordion title="kubectl get models は API に送信する名前を返しません">
  このコマンドが一覧表示するのは、Kubernetes のリソース名 (`metadata.name`) です。推論リクエストにはサービング名 (`spec.name`) を使用する必要があり、以下の **Model ID** 見出しに記載されているのはこのサービング名です。

  ```yaml theme={}
  apiVersion: sambanova.ai/v1alpha1
  kind: Model
  metadata:
    name: gemma-4-31b-it    # Kubernetes resource name – what kubectl get models returns
  spec:
    name: gemma-4-31B-it    # serving name – use this in API requests
  ```

  両者は大文字・小文字や記号のみが異なることが多く、不一致に気づきにくくなっています。上の例では、1 文字だけ異なります。

  デプロイ済みモデルのサービング名を確認するには、次のコマンドを実行します。

  ```shellscript theme={}
  kubectl -n <namespace> describe model <resource-name>
  ```

  モデルには `spec.aliases` を定義することもできます。これは、同じモデルにルーティングされる追加の名前です。`Model` の全フィールドのリファレンスは、[カスタムチェックポイントのデプロイ](/docs/ja/sambastack/service-administration/model-deployment/deploying-models-and-bundles/deploy-custom-checkpoints)をご覧ください。

  音声合成は、名前が一致しないケースの 1 つです。`/v1/audio/speech` には、talker や vocoder のリソース名ではなく `qwen3-tts` を指定してください。詳細は[音声合成](/docs/ja/features/speech)をご覧ください。
</Accordion>

## 対応モデルの一覧

検索またはフィルターでモデルを探し、選択するとそのモデルの完全な構成へ移動できます。

<StackModelExplorer
  models={[
{ id: "Meta-Llama-3.3-70B-Instruct", provider: "Meta", context: "128k", byoc: true, specdec: true, modalities: ["Text"], capabilities: ["Function calling","JSON mode"], use: "Task agent, Tool-calling agent, Text to SQL/Cipher" },
{ id: "Meta-Llama-3.1-8B-Instruct", provider: "Meta", context: "16k", byoc: true, modalities: ["Text"], capabilities: ["Function calling","JSON mode"], use: "Gateway agent, Validation agent" },
{ id: "Meta-Llama-3.1-70B-Instruct", byoc: true, provider: "Meta", context: "32k", specdec: true, modalities: ["Text"], capabilities: ["Function calling","JSON mode"], use: "Task agent, Tool-calling agent" },
{ id: "Meta-Llama-3.1-405B-Instruct", byoc: true, provider: "Meta", context: "16k", specdec: true, modalities: ["Text"], capabilities: ["Function calling","JSON mode"], use: "Task agent, Tool-calling agent, Code generation" },
{ id: "Llama-4-Maverick-17B-128E-Instruct", provider: "Meta", context: "128k", modalities: ["Image","Text"], capabilities: ["Function calling","JSON mode"], use: "Image understanding, Task agent, Tool-calling agent" },
{ id: "MiniMax-M3", provider: "MiniMax", context: "1024k", modalities: ["Image","Text","Video"], capabilities: ["Image and video understanding","Optional thinking mode","Function calling"], use: "Coding agent, Long-context agentic tasks, Image and video understanding" },
{ id: "MiniMax-M2.7", byoc: true, provider: "MiniMax", context: "192k", modalities: ["Text"], capabilities: ["Function calling","Structured output"], use: "Coding agent" },
{ id: "MiniMax-M2.5", byoc: true, provider: "MiniMax", context: "160k", modalities: ["Text"], capabilities: ["Function calling","Structured output"], use: "Coding agent" },
{ id: "Mistral-Large-3-675B-Instruct-2512", byoc: true, provider: "Mistral AI", context: "32k", modalities: ["Text"], capabilities: ["Function calling"], use: "Multilingual instruction following, Task agent, Tool-calling agent" },
{ id: "DeepSeek-R1-0528", byoc: true, provider: "DeepSeek", context: "128k", modalities: ["Text"], capabilities: ["Function calling","JSON mode","Reasoning"], use: "Complex reasoning" },
{ id: "DeepSeek-R1-Distill-Llama-70B", provider: "DeepSeek", context: "128k", byoc: true, specdec: true, modalities: ["Text"], capabilities: ["Reasoning"], use: "Complex reasoning" },
{ id: "DeepSeek-V3-0324", byoc: true, provider: "DeepSeek", context: "128k", modalities: ["Text"], capabilities: ["Function calling","JSON mode"], use: "Main/planner agent, Tool-calling agent" },
{ id: "DeepSeek-V3.1", byoc: true, provider: "DeepSeek", context: "128k", modalities: ["Text"], capabilities: ["Function calling","JSON mode","Reasoning"], use: "Main/planner agent, Tool-calling agent" },
{ id: "DeepSeek-V3.2", byoc: true, provider: "DeepSeek", context: "128k", modalities: ["Text"], capabilities: ["Optional thinking mode","Function calling","JSON mode"], use: "Main/planner agent, Tool-calling agent" },
{ id: "DeepSeek-V3.1-Terminus", byoc: true, provider: "DeepSeek", context: "128k", modalities: ["Text"], capabilities: ["Function calling","JSON mode","Reasoning"], use: "Main/planner agent, Tool-calling agent" },
{ id: "gpt-oss-120b", byoc: true, provider: "OpenAI", context: "128k", modalities: ["Text"], capabilities: ["Reasoning","Function calling","JSON mode","Logit masking","logit_bias sampling parameter"], use: "Main/planner agent, Tool-calling agent" },
{ id: "gpt-oss-20b", byoc: true, provider: "OpenAI", context: "128k", modalities: ["Text"], capabilities: ["Function calling","JSON mode"], use: "Main/planner agent, Tool-calling agent, Reasoning" },
{ id: "Whisper-Large-v3", provider: "OpenAI", context: "—", modalities: ["Audio"], capabilities: [], use: "Automatic speech recognition (ASR), Audio transcription" },
{ id: "gemma-3-27b-it", provider: "Google", context: "128k", modalities: ["Image","Text"], capabilities: ["Image understanding","JSON mode"], use: "Image understanding, Task agent" },
{ id: "gemma-3-12b-it", provider: "Google", context: "128k", modalities: ["Image","Text"], capabilities: ["Image understanding","JSON mode"], use: "Image understanding, Task agent" },
{ id: "gemma-4-31B-it", provider: "Google", context: "256k", modalities: ["Image","Text","Video"], capabilities: ["Image and video understanding","Optional thinking mode","Function calling","JSON mode"], use: "Image and video understanding, Task agent" },
{ id: "Qwen3-235B-A22B-Instruct-2507", byoc: true, provider: "Alibaba Cloud", context: "128k", modalities: ["Text"], capabilities: ["Reasoning"], use: "Agentic planner, Multilingual instruction following" },
{ id: "Qwen3-32B", provider: "Alibaba Cloud", context: "32k", modalities: ["Text"], capabilities: ["Reasoning"], use: "Task agent, Multilingual instruction following" },
{ id: "Qwen3-TTS-Talker", provider: "Alibaba Cloud", context: "4k", modalities: ["Audio","Text"], capabilities: [], use: "Text-to-speech" },
{ id: "Qwen3-TTS-Vocoder", provider: "Alibaba Cloud", context: "4k", modalities: ["Audio"], capabilities: [], use: "Audio synthesis" },
{ id: "Llama-3.3-Swallow-70B-Instruct-v0.4", byoc: true, provider: "Tokyotech-llm", context: "128k", specdec: true, modalities: ["Text"], capabilities: [], use: "Japanese instruction following, Task agent" },
{ id: "E5-Mistral-7B-Instruct", provider: "Other", context: "4k", modalities: ["Embedding"], capabilities: [], use: "Vector storage and retrieval (RAG)" },
]}
/>

## Meta

### Meta-Llama-3.3-70B-Instruct

<StackModelCard provider="Meta" id="Meta-Llama-3.3-70B-Instruct" use="Task agent, Tool-calling agent, Text to SQL/Cipher" modalities={["Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["4K (1, 2, 4, 8, 16, 32)","8K (1, 2, 4, 8, 16, 32)","16K (1, 2, 4)","32K (1, 2, 4)","64K (1, 2, 4)","128K (1)"]} customCheckpoints={true} specDecoding={true} hf="https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct" />

### Meta-Llama-3.1-8B-Instruct

<StackModelCard provider="Meta" id="Meta-Llama-3.1-8B-Instruct" use="Gateway agent, Validation agent" modalities={["Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["4K (1, 2, 4, 8, 16, 32, 64, 128)","8K (1, 2, 4, 8, 16, 32, 64)","16K (1, 2, 4, 8)"]} customCheckpoints={true} hf="https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct" />

### Meta-Llama-3.1-70B-Instruct

<StackModelCard provider="Meta" id="Meta-Llama-3.1-70B-Instruct" use="Task agent, Tool-calling agent" modalities={["Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 2, 4, 8)","16K (1, 2, 4)","32K (1, 4)"]} customCheckpoints={true} specDecoding={true} hf="https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct" />

### Meta-Llama-3.1-405B-Instruct

<StackModelCard provider="Meta" id="Meta-Llama-3.1-405B-Instruct" use="Task agent, Tool-calling agent, Code generation" modalities={["Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["4K (1, 2, 4)","8K (1)","16K (1)"]} customCheckpoints={true} specDecoding={true} hf="https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct" />

### Llama-4-Maverick-17B-128E-Instruct

<StackModelCard provider="Meta" id="Llama-4-Maverick-17B-128E-Instruct" use="Image understanding, Task agent, Tool-calling agent" modalities={["Image","Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1)","16K (1, 2, 4)","32K (1)","64K (1)","128K (1)"]} customCheckpoints={false} hf="https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct" />

## MiniMax

### MiniMax-M3

<StackModelCard provider="MiniMax" id="MiniMax-M3" use="Coding agent, Long-context agentic tasks, Image and video understanding" modalities={["Image","Text","Video"]} capabilities={["Image and video understanding","Optional thinking mode","Function calling"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"],["Responses","/v1/responses","Supported"]]} sequenceLengths={["8K-32K (1, 2, 4)","64K (1, 2, 4)","128K (1, 2, 4)","256K (1, 2, 4)","512K (1)","1M (1)"]} customCheckpoints={false} hf="https://huggingface.co/MiniMaxAI/MiniMax-M3" />

### MiniMax-M2.7

<StackModelCard provider="MiniMax" id="MiniMax-M2.7" use="Coding agent" modalities={["Text"]} capabilities={["Function calling","Structured output"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K-32K (2, 4, 6, 8)","64K (2, 4)","160K (2)","192K (2)"]} customCheckpoints={true} hf="https://huggingface.co/MiniMaxAI/MiniMax-M2.7" />

### MiniMax-M2.5

<StackModelCard provider="MiniMax" id="MiniMax-M2.5" use="Coding agent" modalities={["Text"]} capabilities={["Function calling","Structured output"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K-32K (2, 4, 6, 8)","160K (2)"]} customCheckpoints={true} hf="https://huggingface.co/MiniMaxAI/MiniMax-M2.5" />

## Mistral AI

### Mistral-Large-3-675B-Instruct-2512

<StackModelCard provider="Mistral AI" id="Mistral-Large-3-675B-Instruct-2512" use="Multilingual instruction following, Task agent, Tool-calling agent" modalities={["Text"]} capabilities={["Function calling"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1, 2)","32K (1)"]} customCheckpoints={true} hf="https://huggingface.co/mistralai/Mistral-Large-3-675B-Instruct-2512" />

## DeepSeek

### DeepSeek-R1-0528

<StackModelCard provider="DeepSeek" id="DeepSeek-R1-0528" use="Complex reasoning" modalities={["Text"]} capabilities={["Function calling","JSON mode","Reasoning"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1)","32K (1)","128K (1)"]} customCheckpoints={true} hf="https://huggingface.co/deepseek-ai/DeepSeek-R1-0528" />

### DeepSeek-R1-Distill-Llama-70B

<StackModelCard provider="DeepSeek" id="DeepSeek-R1-Distill-Llama-70B" use="Complex reasoning" modalities={["Text"]} capabilities={["Reasoning"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["4K (2, 4, 8, 16, 32)","8K (2, 4, 8, 16, 32)","16K (1, 2, 4)","32K (1, 2, 4)","64K (1, 2, 4)","128K (1)"]} customCheckpoints={true} specDecoding={true} hf="https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Llama-70B" />

### DeepSeek-V3-0324

<StackModelCard provider="DeepSeek" id="DeepSeek-V3-0324" use="Main/planner agent, Tool-calling agent" modalities={["Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1)","32K (1)","128K (1)"]} customCheckpoints={true} hf="https://huggingface.co/deepseek-ai/DeepSeek-V3-0324" />

### DeepSeek-V3.1

<StackModelCard provider="DeepSeek" id="DeepSeek-V3.1" use="Main/planner agent, Tool-calling agent" modalities={["Text"]} capabilities={["Function calling","JSON mode","Reasoning"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1)","32K (1)","128K (1)"]} customCheckpoints={true} hf="https://huggingface.co/deepseek-ai/DeepSeek-V3.1" />

### DeepSeek-V3.2

<StackModelCard provider="DeepSeek" id="DeepSeek-V3.2" use="Main/planner agent, Tool-calling agent" modalities={["Text"]} capabilities={["Optional thinking mode","Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1)","32K (1)","128K (1)"]} customCheckpoints={true} hf="https://huggingface.co/deepseek-ai/DeepSeek-V3.2" />

### DeepSeek-V3.1-Terminus

<StackModelCard provider="DeepSeek" id="DeepSeek-V3.1-Terminus" use="Main/planner agent, Tool-calling agent" modalities={["Text"]} capabilities={["Function calling","JSON mode","Reasoning"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1)","32K (1)","128K (1)"]} customCheckpoints={true} hf="https://huggingface.co/deepseek-ai/DeepSeek-V3.1-Terminus" />

## OpenAI

### gpt-oss-120b

<StackModelCard provider="OpenAI" id="gpt-oss-120b" use="Main/planner agent, Tool-calling agent" modalities={["Text"]} capabilities={["Reasoning","Function calling","JSON mode","Logit masking","logit_bias sampling parameter"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K-32K (2, 4, 6, 8)","64K (2, 4)","128K (2)"]} customCheckpoints={true} hf="https://huggingface.co/openai/gpt-oss-120b" />

### gpt-oss-20b

<StackModelCard provider="OpenAI" id="gpt-oss-20b" use="Main/planner agent, Tool-calling agent, Reasoning" modalities={["Text"]} capabilities={["Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K-32K (2, 4, 6, 8)","64K (2, 4)","128K (2)"]} customCheckpoints={true} hf="https://huggingface.co/openai/gpt-oss-20b" />

### Whisper-Large-v3

<StackModelCard provider="OpenAI" id="Whisper-Large-v3" use="Automatic speech recognition (ASR), Audio transcription" modalities={["Audio"]} endpoints={[["Translation","/v1/audio/translations","Supported"],["Transcription","/v1/audio/transcriptions","Supported"]]} sequenceLengths={["448 (1, 16, 32)"]} customCheckpoints={false} hf="https://huggingface.co/openai/whisper-large-v3" />

## Google

### gemma-3-27b-it

<StackModelCard provider="Google" id="gemma-3-27b-it" use="Image understanding, Task agent" modalities={["Image","Text"]} capabilities={["Image understanding","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["4K-128K (2, 4, 6, 8)"]} customCheckpoints={false} hf="https://huggingface.co/google/gemma-3-27b-it" />

### gemma-3-12b-it

<StackModelCard provider="Google" id="gemma-3-12b-it" use="Image understanding, Task agent" modalities={["Image","Text"]} capabilities={["Image understanding","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["128K (2, 4, 6, 8)"]} customCheckpoints={false} hf="https://huggingface.co/google/gemma-3-12b-it" />

### gemma-4-31B-it

<StackModelCard provider="Google" id="gemma-4-31B-it" use="Image and video understanding, Task agent" modalities={["Image","Text","Video"]} capabilities={["Image and video understanding","Optional thinking mode","Function calling","JSON mode"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (2, 4, 6, 8)","32K-128K (2, 4, 6, 8)","256K (2)"]} customCheckpoints={false} hf="https://huggingface.co/google/gemma-4-31B-it" />

## Alibaba Cloud

### Qwen3-235B-A22B-Instruct-2507

<StackModelCard provider="Alibaba Cloud" id="Qwen3-235B-A22B-Instruct-2507" use="Agentic planner, Multilingual instruction following" modalities={["Text"]} capabilities={["Reasoning"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["32K (2, 4, 6, 8)","128K (2)"]} customCheckpoints={true} hf="https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507" />

### Qwen3-32B

<StackModelCard provider="Alibaba Cloud" id="Qwen3-32B" use="Task agent, Multilingual instruction following" modalities={["Text"]} capabilities={["Reasoning"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["8K (1, 4)","16K (1)","32K (1, 2)"]} customCheckpoints={false} hf="https://huggingface.co/Qwen/Qwen3-32B" />

### Qwen3-TTS-Talker

<StackModelCard provider="Alibaba Cloud" id="Qwen3-TTS-Talker" use="Text-to-speech" modalities={["Audio","Text"]} endpoints={[["Speech","","Supported"]]} sequenceLengths={["4K (2, 4, 8, 16)"]} customCheckpoints={false} hf="https://huggingface.co/collections/Qwen/qwen3-tts" />

### Qwen3-TTS-Vocoder

<StackModelCard provider="Alibaba Cloud" id="Qwen3-TTS-Vocoder" use="Audio synthesis" modalities={["Audio"]} endpoints={[["Speech","","Supported"]]} sequenceLengths={["Codes length 5, 8, 10, 16 (16)","4K (1)"]} customCheckpoints={false} hf="https://huggingface.co/collections/Qwen/qwen3-tts" />

## Tokyotech-llm

### Llama-3.3-Swallow-70B-Instruct-v0.4

<StackModelCard provider="Tokyotech-llm" id="Llama-3.3-Swallow-70B-Instruct-v0.4" use="Japanese instruction following, Task agent" modalities={["Text"]} endpoints={[["Chat completions","/v1/chat/completions","Supported"]]} sequenceLengths={["4K (1, 2, 4, 8, 16)","8K (1, 2, 4, 8, 16)","16K (1, 2, 4)","32K (1, 2, 4)","64K (1)","128K (1)"]} customCheckpoints={true} specDecoding={true} hf="https://huggingface.co/tokyotech-llm/Llama-3.3-Swallow-70B-Instruct-v0.4" />

## Other

### E5-Mistral-7B-Instruct

<StackModelCard provider="Other" id="E5-Mistral-7B-Instruct" use="Vector storage and retrieval (RAG)" modalities={["Embedding"]} endpoints={[["Embeddings","/v1/embeddings","Supported"]]} sequenceLengths={["4K (1, 4, 8, 16, 32)"]} customCheckpoints={false} hf="https://huggingface.co/intfloat/e5-mistral-7b-instruct" />

## 推奨モデルバンドル

SambaStack のバンドルとは、1 つ以上のモデルを、バッチサイズやシーケンス長などの関連する構成と共にまとめたデプロイパッケージです。単一のモデルは、バンドルを作成せずに、モデルプロファイルと組み合わせて単体でデプロイすることもできます。

例えば、`Meta‑Llama‑3.3‑70B` モデルをバッチサイズ 4、シーケンス長 16K トークンでデプロイする場合、これは 1 つの構成に相当します。一方、バンドルには、同じモデルまたは異なるモデルに対するこのような構成を複数含めることができます。

SambaNova の RDU 技術により、1 つのデプロイメント内に複数のモデルと構成を同時にロードできます。これにより、必要に応じてモデル間やバッチサイズ/シーケンス長のプロファイル間を即座に切り替えられます。デプロイが通常は単一モデルかつ静的である従来の GPU システムとは異なり、SambaStack はマルチモデル・マルチ構成のバンドルをサポートします。このアプローチにより、低レイテンシを維持しながら、より高い効率性、柔軟性、スループットを実現します。

クラスターで利用可能なバンドルは、次のコマンドで確認できます。

```shellscript theme={}
kubectl -n <namespace> get modelbundles
```

以下の表は、SambaStack で現在利用可能なモデルに対する推奨バンドルの一覧です。各エントリは、モデルとその推奨デプロイバンドルの組み合わせを示しています。

<Note>
  以下に記載されたバンドルが推論要件を満たさない場合は、DDR メモリに収まる範囲で、任意のモデルと構成を組み合わせた[カスタムバンドル](/docs/ja/sambastack/service-administration/model-deployment/deploying-models-and-bundles/create-a-custom-bundle)を作成できます。
</Note>

### バンドル構成

以下の表は、各バンドルの構成の詳細です。

| モデル名 | バンドル名 | バンドルの説明 | バンドル構成 |
| - | :- | :- | :- |
| `DeepSeek-R1-0528` / <br />`DeepSeek-V3.1` | deepseek-r1-v31-fp8-8k | <ul><li>中程度のコンテキスト長と低バッチサイズ</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`DeepSeek-R1-0528`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li><li>`DeepSeek-V3.1`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li></ul></details> |
| `DeepSeek-R1-0528` / <br />`DeepSeek-V3-0324` | deepseek-r1-v3-fp8-8k | <ul><li>中程度のコンテキスト長と低バッチサイズ</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`DeepSeek-R1-0528`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li><li>`DeepSeek-V3-0324`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li></ul></details> |
| `DeepSeek-V3-0324` / <br />`DeepSeek-V3.1` | deepseek-v3-v31-fp8-8k | <ul><li>中程度のコンテキスト長と低バッチサイズ</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`DeepSeek-V3-0324`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li><li>`DeepSeek-V3.1`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li></ul></details> |
| `DeepSeek-V3-0324` / <br />`DeepSeek-V3.1-Terminus` | deepseek-v3-v3termi-fp8-8k | <ul><li>中程度のコンテキスト長と低バッチサイズ</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`DeepSeek-V3-0324`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li><li>`DeepSeek-V3.1-Terminus`<ul><li>Seq Length: 8K, BS: 1, 4</li></ul></li></ul></details> |
| `DeepSeek-R1-0528` / <br />`DeepSeek-V3-0324` / <br />`DeepSeek-V3.1` / <br />`DeepSeek-V3.1-Terminus` | deepseek-4in1-fp8-128k | <ul><li>長いコンテキスト長と単一バッチサイズ</li><li>4 つの DeepSeek モデルを 1 つのバンドルに集約</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`DeepSeek-R1-0528`<ul><li>Seq Length: 128K, BS: 1</li></ul></li><li>`DeepSeek-V3-0324`<ul><li>Seq Length: 128K, BS: 1</li></ul></li><li>`DeepSeek-V3.1`<ul><li>Seq Length: 128K, BS: 1</li></ul></li><li>`DeepSeek-V3.1-Terminus`<ul><li>Seq Length: 128K, BS: 1</li></ul></li></ul></details> |
| `E5-Mistral-7B-Instruct` / <br />`Meta-Llama-3.1-8B-Instruct` / <br />`Llama-4-Maverick-17B-128E-Instruct` / <br />`Meta-Llama-3.3-70B-Instruct` / <br />`gpt-oss-120b` | us-agentic-rag-1-1 | <ul><li>小〜中程度のコンテキスト長と多様なバッチサイズ</li><li>`Meta-Llama-3.3-70B` で投機的デコーディングをサポート</li></ul> | <details><summary>View</summary><ul><li>`gpt-oss-120b`<ul><li>Seq Length: 32K, BS: 4</li><li>Seq Length: 64K, BS: 2</li><li>Seq Length: 128K, BS: 2</li></ul></li><li>`Llama-4-Maverick-17B-128E-Instruct`<ul><li>Seq Length: 8K, BS: 1</li><li>Seq Length: 16K, BS: 1</li></ul></li><li>`Meta-Llama-3.3-70B` (Target)/ `Meta-Llama-3.2-1B` (Draft)<ul><li>Seq Length: 4K, BS: 1, 4, 8, 16, 32</li><li>Seq Length: 8K, BS: 1, 4, 8</li><li>Seq Length: 16K, BS: 1, 4</li><li>Seq Length: 32K, BS: 1, 4</li><li>Seq Length: 64K, BS: 1</li><li>Seq Length: 128K, BS: 1</li></ul></li><li>`Meta-Llama-3.1-8B-Instruct`<ul><li>Seq Length: 4K, BS: 1, 4, 16, 32</li><li>Seq Length: 8K, BS: 1, 4, 16, 32</li><li>Seq Length: 16K, BS: 1, 4, 8</li></ul></li><li>`E5-Mistral-7B-Instruct`<ul><li>Seq Length: 4K, BS: 1, 4, 8, 16, 32</li></ul></li></ul></details> |
| `E5-Mistral-7B-Instruct` / <br />`Meta-Llama-3.3-70B` | e5-mistral-70b-64k-128k | <ul><li>長いコンテキスト長と低バッチサイズ</li><li>`Meta-Llama-3.3-70B` で投機的デコーディングをサポート</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`E5-Mistral-7B-Instruct`<ul><li>Seq Length: 4K, BS: 1, 4, 8, 16, 32</li></ul></li><li>`Meta-Llama-3.3-70B` (Target)/ `Meta-Llama-3.2-1B` (Draft)<ul><li>Seq Length: 64K, BS: 1</li><li>Seq Length: 128K, BS: 1</li></ul></li></ul></details> |
| `gemma-3-27b-it` | gemma3-27b-32-128k | `gemma-3-27b-it` の構成のみを含む単一モデルバンドル (homogeneous bundle)。 | <details><summary>View</summary><ul><li>`gemma-3-27b-it`<ul><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 128K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `gemma-3-12b-it` | gemma3-v3 | <ul><li>`gemma-3-12b-it` の構成のみを含む単一モデルバンドル。</li><li>長いコンテキスト長と中程度のバッチサイズ</li></ul> | <details><summary>View</summary><ul><li>`gemma-3-12b-it`<ul><li>Seq Length: 128K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `gemma-4-31B-it` | gemma-4-31b-32-128k | `gemma-4-31B-it` 向けの制約付きデコーディング (constrained decoding) に対応した単一モデルバンドル。テキスト・画像・動画入力に対応。 | <details><summary>View</summary><ul><li>`gemma-4-31B-it`<ul><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 128K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `gemma-4-31B-it` | `gemma-4-31b-32-128-256k` | <ul><li>`gemma-4-31B-it` 向けの制約付きデコーディングに対応した単一モデルバンドル。最大 256K までのコンテキストサポートを追加。すべてのシーケンス長でテキスト・画像・動画入力に対応。</li></ul> | <details><summary>View</summary><ul><li>`gemma-4-31B-it`<ul><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 128K, BS: 2, 4, 6, 8</li><li>Seq Length: 256K, BS: 2</li></ul></li></ul></details> |
| `gemma-4-31B-it`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `gemma-4-31b-mtp-cd-8-32-64-128k` | <ul><li>`gemma-4-31B-it` 向けのマルチトークン予測 (multi-token prediction; MTP) と制約付きデコーディングに対応した単一モデルバンドル。最大 128K のコンテキスト。すべてのシーケンス長でテキスト・画像・動画入力に対応。</li><li>マルチトークン予測はデコードスループットを高めます。モデルが 1 ステップで複数のトークンを予測し、同じパス内で検証します。</li></ul> | <details><summary>View</summary><ul><li>`gemma-4-31B-it`<ul><li>Seq Length: 8K, BS: 2, 4, 6, 8</li><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 64K, BS: 2, 4, 6, 8</li><li>Seq Length: 128K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `gpt-oss-120b` | `cd-dyt-gpt-oss-120b-8-32-64-128k` † | <ul><li>`gpt-oss-120b` 向けの制約付きデコーディングに対応した単一モデルバンドル。</li><li>8K〜128K のコンテキストをカバー。</li><li>構造化出力 (logit マスキング) のサポートを含む。</li></ul> | <details><summary>View</summary><ul><li>`cd-dyt-gpt-oss-120b-8-32-64-128k`<ul><li>Seq Length: 8K, BS: 2, 4, 6, 8</li><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 64K, BS: 2, 4</li><li>Seq Length: 128K, BS: 2</li></ul></li></ul></details> |
| `gpt-oss-20b` | `dyt-gpt-oss-20b-32-64-128k` † | `gpt-oss-20b` 向けの制約付きデコーディングに対応した単一モデルバンドル。 | <details><summary>View</summary><ul><li>`dyt-gpt-oss-20b-32-64-128k`<ul><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 64K, BS: 2, 4</li><li>Seq Length: 128K, BS: 2</li></ul></li></ul></details> |
| `E5-Mistral-7B-Instruct` / <br />`Whisper-Large-v3` / <br />`gemma-3-12b-it` / <br />`gpt-oss-120b` | gpt-gemma-whisper-mistral | <ul><li>Embedding、文字起こし、画像理解、ツール呼び出しエージェントを組み合わせた小〜中程度のコンテキスト長のバンドル</li></ul> | <details><summary>View</summary><p><strong>Models:</strong></p><ul><li>`E5-Mistral-7B-Instruct`<ul><li>Seq Length: 4K, BS: 1</li></ul></li><li>`Whisper-Large-v3`<ul><li>BS: 1</li></ul></li><li>`gemma-3-12b-it`<ul><li>Seq Length: 128K, BS: 2, 4, 6, 8</li></ul></li><li>`gpt-oss-120b`<ul><li>Seq Length: 8K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `Llama-4-Maverick-17B-128E-Instruct` | llama-4-medium-8-16-32-64-128k | <ul><li>`Llama-4-Maverick-17B-128E-Instruct` の構成のみを含む単一モデルバンドル。</li><li>小〜長いコンテキスト長と低バッチサイズ</li></ul> | <details><summary>View</summary><ul><li>`Llama-4-Maverick-17B-128E-Instruct`<ul><li>Seq Length: 8K, BS: 1</li><li>Seq Length: 16K, BS: 1</li><li>Seq Length: 32K, BS: 1</li><li>Seq Length: 64K, BS: 1</li><li>Seq Length: 128K, BS: 1</li></ul></li></ul></details> |
| `Llama-4-Maverick-17B-128E-Instruct` | llama-4-medium-ss-16k-bs24 | `Llama-4-Maverick-17B-128E-Instruct` の構成のみを含む単一モデルバンドル。中程度のコンテキスト長とより大きいバッチサイズ。 | <details><summary>View</summary><ul><li>`Llama-4-Maverick-17B-128E-Instruct`<ul><li>Seq Length: 16K, BS: 2, 4</li></ul></li></ul></details> |
| `Meta-Llama-3.3-70B-Instruct` | 70b-3dot3-ss-4-8-16-32-64-128k | <ul><li>中〜長のコンテキスト長と低バッチサイズ</li></ul> | <details><summary>View</summary><p><strong>Target Models:</strong></p><ul><li>`Meta-Llama-3.3-70B-Instruct`<ul><li>Seq Length: 4K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 8K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 16K, BS: 1, 2, 4</li><li>Seq Length: 32K, BS: 1, 2, 4</li><li>Seq Length: 64K, BS: 1, 2, 4</li><li>Seq Length: 128K, BS: 1</li></ul></li></ul><p><strong>Draft Models:</strong></p><ul><li>`Meta-Llama-3.2-1B-Instruct`<ul><li>Seq Length: 4K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 8K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 16K, BS: 1, 2, 4</li><li>Seq Length: 32K, BS: 1, 2, 4</li><li>Seq Length: 64K, BS: 1, 2, 4; private: true</li><li>Seq Length: 128K, BS: 1; private: true</li></ul></li></ul></details> |
| `Meta-Llama-3.3-70B-Instruct` / <br />`Whisper-Large-v3` | 70b-3dot3-ss-full-whisper | <ul><li>全レンジのコンテキスト長と低〜中程度のバッチサイズに加え、Whisper による文字起こしに対応</li><li>`Meta-Llama-3.3-70B-Instruct` で投機的デコーディングをサポート</li></ul> | <details><summary>View</summary><p><strong>Target Models:</strong></p><ul><li>`Meta-Llama-3.3-70B-Instruct`<ul><li>Seq Length: 4K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 8K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 16K, BS: 1, 2, 4</li><li>Seq Length: 32K, BS: 1, 2, 4</li><li>Seq Length: 64K, BS: 1, 2, 4</li><li>Seq Length: 128K, BS: 1</li></ul></li><li>`Whisper-Large-v3`<ul><li>BS: 1, 16, 32</li></ul></li></ul><p><strong>Draft Models:</strong></p><ul><li>`Meta-Llama-3.2-1B-Instruct`<ul><li>Seq Length: 4K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 8K, BS: 2, 4, 8, 16, 32</li><li>Seq Length: 16K, BS: 1, 2, 4</li><li>Seq Length: 32K, BS: 1, 2, 4</li><li>Seq Length: 64K, BS: 1, 2, 4</li><li>Seq Length: 128K, BS: 1</li></ul></li></ul></details> |
| `MiniMax-M2.5` | <ul><li>dyt-minimax-m2p5-32k</li><li>dyt-minimax-m2p5-32-160k</li></ul> | <ul><li>`MiniMax-M2.5` の構成のみを含む単一モデルバンドル。</li><li>dyt-minimax-m2p5-32k は、中程度のシーケンス長と高バッチング向き。</li><li>dyt-minimax-m2p5-32-160k は、より長いシーケンス長と低バッチング向き</li></ul> | <details><summary>View</summary><ul><li>dyt-minimax-m2p5-32k<ul><li>Seq Length: 4K-32K, BS: 2, 4, 6, 8</li></ul></li></ul><ul><li>dyt-minimax-m2p5-32-160k<ul><li>Seq Length: 32K, BS: 2</li><li>Seq Length: 160K, BS: 2</li></ul></li></ul></details> |
| `MiniMax-M2.7` | dyt-minimax-m2p7-32k-v2 | `MiniMax-M2.7` の構成のみを含む単一モデルバンドル。中程度のコンテキスト長と高バッチング。 | <details><summary>View</summary><ul><li>`MiniMax-M2.7`<ul><li>Seq Length: 8K-32K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `MiniMax-M2.7` | dyt-minimax-m2p7-32-160-192k | `MiniMax-M2.7` の構成のみを含む単一モデルバンドル。より長いシーケンス長と低バッチング向き。 | <details><summary>View</summary><ul><li>`MiniMax-M2.7`<ul><li>Seq Length: 8K-32K, BS: 2, 4, 6, 8</li><li>Seq Length: 160K-192K, BS: 2</li></ul></li></ul></details> |
| `MiniMax-M2.7` | dyt-minimax-m2p7-32-64-160-192k-pc | `MiniMax-M2.7` 向けの Prompt Caching (プロンプトキャッシング) に対応した単一モデルバンドル。 | <details><summary>View</summary><ul><li>`MiniMax-M2.7`<ul><li>Seq Length: 8K-32K, BS: 2, 4, 6, 8</li><li>Seq Length: 64K, BS: 2, 4</li><li>Seq Length: 160K-192K, BS: 2</li></ul></li></ul></details> |
| `MiniMax-M2.7` | dyt-minimax-m2p7-32-64-160-192k | <ul><li>`MiniMax-M2.7` の構成のみを含む単一モデルバンドル。`dyt-minimax-m2p7-32-64-160-192k-pc` と同じコンテキスト範囲で、Prompt Caching なし。</li></ul> | <details><summary>View</summary><ul><li>`MiniMax-M2.7`<ul><li>Seq Length: 8K-32K, BS: 2, 4, 6, 8</li><li>Seq Length: 64K, BS: 2</li><li>Seq Length: 160K-192K, BS: 2</li></ul></li></ul></details> |
| `MiniMax-M2.7` | dyt-minimax-m2p7-32k-pc | `MiniMax-M2.7` 向けの Prompt Caching に対応した単一モデルバンドル。中程度のコンテキスト長と高バッチング。 | <details><summary>View</summary><ul><li>`MiniMax-M2.7`<ul><li>Seq Length: 8K-32K, BS: 2, 4, 6, 8</li></ul></li></ul></details> |
| `MiniMax-M3`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | minimax-m3-32k | `MiniMax-M3` の構成のみを含む単一モデルバンドル。中程度のコンテキスト長。テキスト・画像・動画入力に対応。 | <details><summary>View</summary><ul><li>`MiniMax-M3`<ul><li>Seq Length: 8K-32K, BS: 1, 2, 4</li></ul></li></ul></details> |
| `MiniMax-M3`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | minimax-m3-512k | <ul><li>`MiniMax-M3` の構成のみを含む単一モデルバンドル。</li><li>最大 512K のコンテキスト。すべてのシーケンス長でテキスト・画像・動画入力に対応。</li></ul> | <details><summary>View</summary><ul><li>`MiniMax-M3`<ul><li>Seq Length: 8K-32K, BS: 1, 2, 4</li><li>Seq Length: 64K, BS: 1, 2, 4</li><li>Seq Length: 128K, BS: 1, 2, 4</li><li>Seq Length: 256K, BS: 1, 2, 4</li><li>Seq Length: 512K, BS: 1</li></ul></li></ul></details> |
| `MiniMax-M3`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | minimax-m3-32-64-128-256-512k-1m | <ul><li>`MiniMax-M3` の構成のみを含む単一モデルバンドル。</li><li>最大 1M のコンテキスト。すべてのシーケンス長でテキスト・画像・動画入力に対応。</li></ul> | <details><summary>View</summary><ul><li>`MiniMax-M3`<ul><li>Seq Length: 8K-32K, BS: 1, 2, 4</li><li>Seq Length: 64K, BS: 1, 2, 4</li><li>Seq Length: 128K, BS: 1, 2, 4</li><li>Seq Length: 256K, BS: 1, 2, 4</li><li>Seq Length: 512K, BS: 1</li><li>Seq Length: 1M, BS: 1</li></ul></li></ul></details> |
| `MiniMax-M3`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `minimax-m3-acb-prefix-caching` | <ul><li>`MiniMax-M3` 向けの Prompt Caching、連続バッチング (continuous batching)、制約付きデコーディングに対応した単一モデルバンドル。最大 1M のコンテキスト。</li><li>テキスト入力のみ。他の `MiniMax-M3` バンドルは、テキスト・画像・動画に対応します。</li></ul> | <details><summary>View</summary><ul><li>`MiniMax-M3`<ul><li>Seq Length: 1M, BS: 2</li></ul></li></ul></details> |
| `Mistral-Large-3-675B-Instruct-2512`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | mistral-large-3-fp8-8k | <ul><li>`Mistral-Large-3-675B-Instruct-2512` の構成のみを含む単一モデルバンドル。</li><li>プレビューモデル – テキストのみ。</li></ul> | <details><summary>View</summary><ul><li>`Mistral-Large-3-675B-Instruct-2512`<ul><li>Seq Length: 8K, BS: 1</li></ul></li></ul></details> |
| `Mistral-Large-3-675B-Instruct-2512`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | mistral-large-3-fp8-8-16-32k | <ul><li>`Mistral-Large-3-675B-Instruct-2512` の構成のみを含む単一モデルバンドル。</li><li>`mistral-large-3-fp8-8k` より広いコンテキスト範囲。プレビューモデル – テキストのみ。</li></ul> | <details><summary>View</summary><ul><li>`Mistral-Large-3-675B-Instruct-2512`<ul><li>Seq Length: 8K, BS: 1, 4</li><li>Seq Length: 16K, BS: 1, 2</li><li>Seq Length: 32K, BS: 1</li></ul></li></ul></details> |
| `Qwen3-235B-A22B-Instruct-2507` | dyt-qwen3-235b-32-128k | `Qwen3-235B-A22B-Instruct-2507` の構成のみを含む単一モデルバンドル。 | <details><summary>View</summary><ul><li>`Qwen3-235B-A22B-Instruct-2507`<ul><li>Seq Length: 32K, BS: 2, 4, 6, 8</li><li>Seq Length: 128K, BS: 2</li></ul></li></ul></details> |
| `Whisper-Large-v3` / <br />`Qwen3-32B` / <br />`E5-Mistral-7B-Instruct` | qwen3-32b-whisper-e5-mistral | <ul><li>小〜中程度のコンテキスト長と多様なバッチサイズ</li></ul> | <details><summary>View</summary><ul><li>`E5-Mistral-7B-Instruct`<ul><li>Seq Length: 4K, BS: 1, 4, 8, 16, 32</li></ul></li><li>`Qwen3-32B`<ul><li>Seq Length: 8K, BS: 1, 4</li><li>Seq Length: 16K, BS: 1</li><li>Seq Length: 32K, BS: 1, 2</li></ul></li><li>`Whisper-Large-v3`<ul><li>BS: 1, 16, 32</li></ul></li></ul></details> |
| `Qwen3-32B` / <br />`Meta-Llama-3.1-405B-Instruct` | qwen3-32b-llama405b-s-m | <ul><li>小〜中程度のコンテキスト長</li><li>`Meta-Llama-3.1-405B-Instruct` で投機的デコーディングをサポート</li></ul> | <details><summary>View</summary><p><strong>Target Models:</strong></p><ul><li>`Meta-Llama-3.1-405B-Instruct`<ul><li>Seq Length: 4K, BS: 1, 2, 4</li><li>Seq Length: 8K, BS: 1</li><li>Seq Length: 16K, BS: 1</li></ul></li></ul><p><strong>Draft Models:</strong></p><ul><li>`Meta-Llama-3.1-8B-Instruct`<ul><li>Seq Length: 16K, BS: 1</li></ul></li><li>`Meta-Llama-3.2-3B-Instruct`<ul><li>Seq Length: 4K, BS: 1, 2, 4</li><li>Seq Length: 8K, BS: 1</li></ul></li></ul><p><strong>Routable Models:</strong></p><ul><li>`Qwen3-32B`<ul><li>Seq Length: 8K, BS: 1, 4</li><li>Seq Length: 16K, BS: 1</li><li>Seq Length: 32K, BS: 1</li></ul></li></ul></details> |
| `Qwen3-TTS-Talker`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | qwen3-tts-talker | `Qwen3-TTS-Talker` の構成のみを含む単一モデルバンドル。 | <details><summary>View</summary><ul><li>`Qwen3-TTS-Talker`<ul><li>Seq Length: 4K, BS: 2, 4, 8, 16</li></ul></li></ul></details> |
| `Qwen3-TTS-Vocoder`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | qwen3-tts-vocoder | `Qwen3-TTS-Vocoder` の構成のみを含む単一モデルバンドル。 | <details><summary>View</summary><ul><li>`Qwen3-TTS-Vocoder`<ul><li>Codes length: 5, 8, 10, 16, BS: 16</li></ul></li></ul></details> |

† `cd-dyt-gpt-oss-120b-8-32-64-128k` は 8K〜128K、`dyt-gpt-oss-20b-32-64-128k` は 32K〜128K のシーケンス長をサポートします。いずれかのモデルでより短いコンテキスト長 (4K〜16K) が必要な場合は、SambaNova の担当者にお問い合わせください。

### モデルごとの推奨バンドル

このセクションでは、モデルごとに、一般的な用途に適した推奨バンドルと、コンテキスト長・バッチサイズ・対応モダリティのトレードオフがある代替バンドルを記載します。各バンドルのシーケンス長/バッチサイズの詳細は、上記の[バンドル構成](#バンドル構成)をご参照ください。

| 開発元 | モデル | 推奨バンドル | 代替バンドル |
| - | - | - | - |
| Meta | `Meta-Llama-3.3-70B-Instruct` | `70b-3dot3-ss-4-8-16-32-64-128k` | `70b-3dot3-ss-full-whisper`, `us-agentic-rag-1-1`, `e5-mistral-70b-64k-128k` |
| Meta | `Meta-Llama-3.1-8B-Instruct` | `us-agentic-rag-1-1` | `qwen3-32b-llama405b-s-m` |
| Meta | `Meta-Llama-3.1-405B-Instruct` | `qwen3-32b-llama405b-s-m` | — |
| Meta | `Llama-4-Maverick-17B-128E-Instruct` | `llama-4-medium-8-16-32-64-128k` | `llama-4-medium-ss-16k-bs24`, `us-agentic-rag-1-1` |
| MiniMax | `MiniMax-M3`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `minimax-m3-32-64-128-256-512k-1m` (最大 1M のコンテキスト) | `minimax-m3-512k` (最大 512K のコンテキスト), `minimax-m3-32k`, `minimax-m3-acb-prefix-caching` ([Prompt Caching](/docs/ja/sambastack/service-administration/performance/prompt-caching) を追加。テキストのみ) |
| MiniMax | `MiniMax-M2.7` | `dyt-minimax-m2p7-32-64-160-192k-pc` ([Prompt Caching](/docs/ja/sambastack/service-administration/performance/prompt-caching) を追加) | `dyt-minimax-m2p7-32k-pc`, `dyt-minimax-m2p7-32-64-160-192k`, `dyt-minimax-m2p7-32-160-192k`, `dyt-minimax-m2p7-32k-v2` |
| MiniMax | `MiniMax-M2.5` | `dyt-minimax-m2p5-32-160k` | `dyt-minimax-m2p5-32k` |
| Mistral AI | `Mistral-Large-3-675B-Instruct-2512`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `mistral-large-3-fp8-8-16-32k` | `mistral-large-3-fp8-8k` |
| DeepSeek | `DeepSeek-R1-0528` | `deepseek-4in1-fp8-128k` (長いコンテキスト長) | `deepseek-r1-v3-fp8-8k`, `deepseek-r1-v31-fp8-8k` |
| DeepSeek | `DeepSeek-V3-0324` | `deepseek-4in1-fp8-128k` (長いコンテキスト長) | `deepseek-r1-v3-fp8-8k`, `deepseek-v3-v31-fp8-8k`, `deepseek-v3-v3termi-fp8-8k` |
| DeepSeek | `DeepSeek-V3.1` | `deepseek-4in1-fp8-128k` (長いコンテキスト長) | `deepseek-r1-v31-fp8-8k`, `deepseek-v3-v31-fp8-8k` |
| DeepSeek | `DeepSeek-V3.1-Terminus` | `deepseek-4in1-fp8-128k` (長いコンテキスト長) | `deepseek-v3-v3termi-fp8-8k` |
| OpenAI | `gpt-oss-120b` | `us-agentic-rag-1-1` | `cd-dyt-gpt-oss-120b-8-32-64-128k`, `gpt-gemma-whisper-mistral` |
| OpenAI | `gpt-oss-20b`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `dyt-gpt-oss-20b-32-64-128k` | — |
| OpenAI | `Whisper-Large-v3` | `qwen3-32b-whisper-e5-mistral` | `70b-3dot3-ss-full-whisper` (既知の問題: デフォルトの起動時間内にロードが完了しません), `gpt-gemma-whisper-mistral` |
| Google | `gemma-3-27b-it`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `gemma3-27b-32-128k` | — |
| Google | `gemma-3-12b-it`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `gemma3-v3` | `gpt-gemma-whisper-mistral` |
| Google | `gemma-4-31B-it` | `gemma-4-31b-32-128-256k` (256K コンテキストを追加。テキスト・画像・動画に対応)、`gemma-4-31b-32-128k` (最大 128K のコンテキスト。テキスト・画像・動画に対応、より高いスループット)。どちらのバンドルも、すべてのシーケンス長で制約付きデコーディングに対応しています。両者は機能ではなく、コンテキスト長とスループットで選択してください。 | `gemma-4-31b-mtp-cd-8-32-64-128k` (Preview。マルチトークン予測、最大 128K のコンテキスト) |
| Alibaba Cloud | `Qwen3-235B-A22B-Instruct-2507` | `dyt-qwen3-235b-32-128k` | — |
| Alibaba Cloud | `Qwen3-32B` | `qwen3-32b-whisper-e5-mistral` | `qwen3-32b-llama405b-s-m` |
| Alibaba Cloud | `Qwen3-TTS-Talker`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `qwen3-tts-talker` | — |
| Alibaba Cloud | `Qwen3-TTS-Vocoder`<sup style={{background:"#4338ca",color:"#ffffff",fontWeight:700,fontSize:"0.62em",letterSpacing:"0.03em",padding:"1px 4px",borderRadius:"3px",marginLeft:"3px"}}>PREVIEW</sup> | `qwen3-tts-vocoder` | — |
| Other | `E5-Mistral-7B-Instruct` | `us-agentic-rag-1-1` | `e5-mistral-70b-64k-128k`, `qwen3-32b-whisper-e5-mistral`, `gpt-gemma-whisper-mistral` |
