GPT-6.1 Sol
GPT-6.1 Sol delivers near-Astra intelligence on agentic coding, computer use and professional work at one-fifth of Astra's price.
Upgrade to GPT-6 Sol matching Astra on DeepSWE v1.1 at roughly one-fifth the cost; $2 input and $10 output per million tokens, $0.10 cached input. Treated as Critical for cyber and High for bio under the Preparedness Framework, with Astra's safeguards. Supports Multi-agent beta. OpenAI's Navier-Stokes post says a more capable internal model began training 2026-08-28.
- Date
- Tuesday, 29 September 2026
- Lab
- OpenAI
- Kind
- model
- Access
- closed API
- Price
- $2 input / $10 output per 1M tokens
Figures
| Measure | Value | Measured by |
|---|---|---|
| API price (input / output per 1M tokens) | $2 / $10 cached input $0.10; Astra is $10 / $50 | company |
Sol 6.1 is an upgrade of GPT-6 Sol, not a new pretrain; DevDay 2026 added Ultrafast for Astra and then 6.1 Sol.
Sources
- openai.com/index/introducing-gpt-6-1-sol/
- deploymentsafety.openai.com/gpt-6-1-sol/
- simonwillison.net/2026/Sep/29/hn-49898129/
- developers.openai.com/api/docs/changelog
This record was checked against its sources on 6 October 2026. How we check