GPT-6.1 Sol: OpenAI's $2/$10 Model From DevDay
OpenAI released GPT-6.1 Sol at DevDay on 29 September at $2 and $10 per million tokens. The claims, the caveats and what to check before you switch.
OpenAI released GPT-6.1 Sol at its DevDay event on 29 September 2026. It is priced at $2 per million input tokens and $10 per million output tokens, with cached input at $0.10 per million, and OpenAI says it nearly matches GPT-6 Astra on agentic coding, computer use and professional work at about a fifth of Astra's token prices. The API name is gpt-6.1-sol.
What GPT-6.1 Sol costs and where you can use it
The standard API prices, per The Next Web's launch coverage:
Input: $2 per million tokens
Cached input: $0.10 per million tokens
Output: $10 per million tokens
On availability, GPT-6.1 Sol is open from launch day to Plus, Pro, Business, Enterprise and Edu users in ChatGPT Work and in Codex, and to developers through the API. It is not yet in the main ChatGPT Chat experience. If your users live in Chat, they cannot reach it yet.
It follows GPT-6 Sol, which arrived on 22 September alongside the cheaper Luna model. Our write-up of those launches covers the GPT-6 Sol and Luna price cuts.
The benchmark claims, and who ran them
The figures in the launch coverage come from OpenAI's own reporting:
DeepSWE v1.1 (coding): matches Astra at one fifth of the cost, and beats GPT-6 Sol by 6.4 percentage points
AutomationBench: 2.2 points above Anthropic's Opus 5.5 at about a third of the cost
OSWorld 2.0 (computer use): within 2.1 points of Astra at one seventh of the task cost
Terminal-Bench Science: $5.47 per task on average at maximum effort, against $23.21 for Opus 5.5 and $23.80 for Astra
Factual errors at low reasoning effort: down from 11.4% to 7.7%
Two points on reading these. First, these are numbers OpenAI reported for its own model, and the comparisons pick the competitor and the effort setting. That does not make them wrong, but it makes them a claim to reproduce rather than a fact to plan around. Second, the cost figures are per task, not per token. Per-token prices tell you the rate; per-task cost also depends on how many tokens the model spends, which is the same gap we explain in what token efficiency means.
The model that did not ship
The timing matters. OpenAI cancelled the October launch of GPT-6.1 Astra after safety tests, a story we cover in its own post. GPT-6.1 Sol is the cheaper line that did ship. If you were waiting on Astra, treat Sol as the available option and Astra as unavailable.
How to test it before switching
Pull 20 to 30 real tasks from your logs, not synthetic ones.
Run them on your current model and on gpt-6.1-sol at the reasoning effort you would actually use.
Record tokens, tool calls, cost and pass rate per task. Compare cost per completed task.
Check the failure cases by hand. A cheaper model that fails quietly costs more than an expensive one that fails loudly.
If you already compare models regularly, how to know when to upgrade to a newer AI model has the decision rules, and how to keep up with AI news covers pacing yourself through weeks like this one.
FAQ
How much does GPT-6.1 Sol cost?
$2 per million input tokens, $0.10 per million cached input tokens and $10 per million output tokens on the standard API.
Is GPT-6.1 Sol available in ChatGPT?
It is available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users. It is not yet available in the main Chat experience.
What is the GPT-6.1 Sol API model name?
gpt-6.1-sol.
Is GPT-6.1 Sol better than Opus 5.5?
OpenAI reports it ahead on AutomationBench at lower cost. Those are vendor-run comparisons, so test on your own tasks.
How did this land?
About the author
Senior Editor, AI & Product
Cecilia leads the Swarmz editorial desk. She has spent a decade turning complex AI and product topics into writing people actually finish, and she owns the blog's quality bar.
