ScaledThought
MiMo-V2.6-ProXiaomi · Sep 22 · AA Index 46DeepSeek-V4.1-FlashDeepSeek · Sep 10 · AA Index 39K2 Horizon 375BIFM · Sep 03 · AA Index 31GLM-5.3Z.ai · Aug 29 · AA Index 45Hy4 previewTencent · Aug 28 · not yet scoredGLM-5.3-FlashZ.ai · Aug 26 · AA Index 42Qwen3.8-27BAlibaba · Aug 14 · AA Index 34DeepSeek V4 ProDeepSeek · Aug 13 · AA Index 36MiMo-V2.6-ProXiaomi · Sep 22 · AA Index 46DeepSeek-V4.1-FlashDeepSeek · Sep 10 · AA Index 39K2 Horizon 375BIFM · Sep 03 · AA Index 31GLM-5.3Z.ai · Aug 29 · AA Index 45Hy4 previewTencent · Aug 28 · not yet scoredGLM-5.3-FlashZ.ai · Aug 26 · AA Index 42Qwen3.8-27BAlibaba · Aug 14 · AA Index 34DeepSeek V4 ProDeepSeek · Aug 13 · AA Index 36

Open models,
the day
they ship.

  • 01Latest open frontier modelsKimi, GLM, Qwen, DeepSeek, MiMo. Live the day weights drop.
  • 02Free to startNo card. No forms. No sales call.
  • 03Run entirely by your agentSetup, keys, models, budgets. There is no dashboard.
>

01 / Day zero

Weights drop. You're live before the thread trends.

When a lab publishes open weights, we're already pulling them. Quantization, kernels and evals run in parallel. If we're not live within 24h, your month is free.
Arrivals · open weights · AA Intelligence IndexUTC --:--:--
ModelWeightsAA idxStatus
Recent open-weight model arrivals
DeepSeek-V4.1-FlashDeepSeekSep 10 39LIVE
K2 Horizon 375BIFMSep 03 31LIVE
GLM-5.3Z.aiAug 29 45LIVE
Hy4 previewTencentAug 28 --LIVE
GLM-5.3-FlashZ.aiAug 26 42LIVE

02 / Benchmarks

Know which model wins before you call it.

Every model we serve comes with independent benchmark scores, speed and price, on day one. Your agent reads the same numbers through the API and picks for you.

Compare all models→

Intelligence Index: Artificial Analysis Intelligence Index v4.3.2, as of Sep 23, 2026. GPQA Diamond: lab model cards. Source ↗

  1. 0146
  2. 0245
  3. 03
    Kimi K3Moonshot AI
    44
  4. 0442
  5. 0540
  6. 0639
  7. 0736

03 / Router

model: "auto"

Set a quality bar once. For every request, the router picks the cheapest open model that clears it, and tells you which one it chose and why.

  • Label 40,000 support tickets→GLM-5.3-FlashShort input, one-word output. The fast model clears the bar.
  • Quick answer inside a chat app→Qwen3.8-27BSmall dense model, lowest latency.
  • Summarize a 300-page contract→DeepSeek-V4.1-FlashLong context with a small memory footprint.
  • Refactor a 2,000-line module→Kimi K3Hard, long-horizon coding. Worth the bigger model.

One setting

quality_floor: 0.85. Raise it for hard work, lower it for bulk jobs.

Pin any time

Name a model instead of auto and it never changes under you.

Every answer explains itself

x-st-routed-to and x-st-reason on every response.

04 / Headless

Works with your agent.

No dashboard to learn. Your agent signs up, gets a key and ships, in whatever tool you work in.

Claude Code / Codex / Cursor / any MCP client / the OpenAI SDK / plain curl

Your agent gets tools for accounts, keys, models and usage.

claude mcp add --transport http scaledthought \
  https://api.us.scaledthought.com/mcp \
  --header "Authorization: Bearer $ST_API_KEY"

One command: free account, key in .env, newest models listed.

npx scaledthought@latest init --region us

OpenAI-compatible. Change the base URL and keep your code.

from openai import OpenAI
client = OpenAI(
  base_url="https://api.us.scaledthought.com/v1",
  api_key=os.environ["ST_API_KEY"],
)

05 / Data residency

Your prompts never leave the country you pick.

Choose US or Canada when you create a key. Inference, logs and billing data stay there. Nothing is retained by default, and nothing is ever used for training.

US

United Statesregion: "us"

us-east · us-central

CA

Canadaregion: "ca"

ca-central · ca-west

✓Zero data retention by default

✓Never used for training

✓Region pinned per API key

Security and residency details→

06 / Built for the switch

Moving to open models should be boring.

01

Shadow mode

Mirror 1% of your closed-model traffic. Get a per-task report on quality and cost before you switch.

02

Your evals, not ours

Upload 50 real examples. Every model gets scored on your work, and the router tunes to your pass rate.

03

Hard budgets for agents

Per-key spend caps and daily limits. An agent can hold a key without holding your card.

04

Models never vanish

Pinned versions stay up 12 months after a successor ships.

05

Tool calls that just work

Every model's tool-call dialect normalized to one format, with strict JSON output.

06

Your fine-tunes stay yours

Bring LoRAs, download weights any time. Open should stay open.

Your agent
already knows how.

Opens your agent with the prompt typed. You press Enter.