OpenAI releases GPT-6.1 Sol, claiming near-Astra results at a fifth of the token price

OpenAI shipped GPT-6.1 Sol in its API, Codex and ChatGPT Work, claiming near-Astra coding and agent results at one-fifth of Astra's token prices.

OpenAI has released GPT-6.1 Sol, an upgrade to GPT-6 Sol. OpenAI says it nearly matches GPT-6 Astra on agentic coding, computer use and professional work at one-fifth of Astra's standard token prices. If that holds in your workloads, it could become the default mid-tier model for API agents. For now, every benchmark comes from OpenAI.

The model is live in the API as gpt-6.1-sol at $2 per million input tokens, $10 per million output tokens and $0.10 per million cached input tokens. Kingy AI notes that's half of GPT-6 Sol's cached rate. The New Stack reports the same rates. Kingy AI lists Astra at $10 input and $50 output, and notes that per-token ratios don't guarantee the same saving per completed task. In ChatGPT, the model is available to Plus, Pro, Business, Enterprise and Edu users in Work and Codex, but not yet in Chat. OpenAI says an Ultrafast Codex variant is coming "in the coming days."

OpenAI's own numbers show the model matching Astra on DeepSWE v1.1, coming within 2.1 points of Astra on OSWorld 2.0 offline, and trailing Astra on Terminal-Bench Science (57.0% vs 68.1%; Astra's score is from OpenAI, Sol's from Kingy AI). OpenAI still recommends Astra for the hardest scientific tasks.

Migration isn't a pure model-name swap. DataCamp notes that none and minimal reasoning effort aren't supported, and that tool calling requires the Responses API. DataCamp also flags a cliff: prompts over 272,000 input tokens are billed at 2x the input and cache rates and 1.5x the output rate for the whole request, so watch retrieval-heavy runs. Before you switch production defaults, replay a fixed task set.

Sources