All models

Stealth: Gestalt

Coming soonProvisional ID

A stealth model that tops a design benchmark judged by professional designers, ahead of every named model and the human reference designs. Its maker, model ID and release date are unknown. If it ships and lands on TeamoRouter, your key works as is: just switch the model ID. For design and front-end work today, use Claude Fable 5.2.

Pay as you go from $5 — card, Alipay or USDT.

Status

Stealth (maker unknown)

Official price (in / out)

Not announced

Context

Not announced

Released

Not announced

What we know

  • No lab has claimed Gestalt. Contra Labs calls it "a high-quality design model currently in stealth". Its maker, model ID, price and release date are all unknown, and it may never come as a public API.
  • On Contra's Frontier Design Benchmark, launched November 19, 2026, Gestalt is #1 at about 1,930 Elo. Claude Fable 5.2, the best named model, is at about 1,714.
  • The benchmark uses real client briefs from Contra; professional designers vote blind between two websites. When designers saw a Gestalt site next to a human-made reference design, they picked Gestalt 20 times out of 28.
  • The benchmark is Contra's own, and Gestalt's results haven't been checked independently.

Until it's out, use these three

Same key, just switch the model ID.

Claude Fable 5.2

claude-fable-5-2

The highest-ranked model you can call today on the same benchmark (#3, behind Gestalt and the human designers).

Model page →Quickstart →

Claude Opus 5.5

claude-opus-5-5

Close behind Fable 5.2 on the same benchmark at half its list price.

Model page →Quickstart →

GPT 6.1 Sol

gpt-6.1-sol

OpenAI's best entry on the same benchmark, at a lower list price.

Model page →Quickstart →

Background and sources

  • Posts on X guess at different labs, Anthropic among them. None offers evidence, and Contra hasn't said.

Sources

Full write-up: Gestalt