OpenAI releases GPT‑6.1 Sol at a fifth of Astra's price
OpenAI says GPT-6.1 Sol nearly matches GPT-6 Astra on coding and computer use for $2 per million input tokens, with cached input at $0.10.
OpenAI released GPT-6.1 Sol on Tuesday, describing it as "near-Astra intelligence for a fifth of the price." The Next Web reports the API model, gpt-6.1-sol, costs $2 per million input tokens and $10 per million output tokens. Cached input costs $0.10, which OpenAI says is 95% less than standard input pricing.
https://x.com/OpenAI/status/2104986129686741046
It is available to Plus, Pro, Business, Enterprise and Edu users in ChatGPT Work and Codex, according to OpenAI's posts. The Next Web reports it is not yet in the main ChatGPT chat. Sam Altman called it an "extremely capable model" on X. Every performance figure below comes from OpenAI, and independent testing is only starting.
OpenAI says Sol matches Astra on DeepSWE v1.1 and lands within 2.1 points of it on OSWorld 2.0 at one-seventh the cost per task, according to The Next Web. It also reports Sol scoring 2.2 points above Anthropic's Opus 5.5 on AutomationBench. On safety tests, OpenAI says Sol tried to bypass restrictions in 23.5% of cases, against 64.4% for GPT-6 Sol.
Outside numbers arrived within hours. Artificial Analysis lists the model at 52 on its Intelligence Index, tenth of 221 models, with a 1 million token context window. Developer Theo, citing those figures, said Sol performs at about Opus 5.5 Medium levels for under a third of the price. He later said his own Terminal Bench 4 runs, made with Codex, looked far better than Artificial Analysis's harness produced.
https://x.com/theo/status/2105005625726099473