GPT-6.1 Sol
GPT-6.1 Sol is OpenAI's DevDay 2026 model for coding and agent work. Its API ID is gpt-6.1-sol, with 1,050,000 tokens of context and up to 128,000 output tokens. It is available through the API and eligible paid ChatGPT Work and Codex plans; OpenAI says Chat availability is still to come.
Read the official API model contract ↗Why it matters
The model offers an updated Sol option without changing the standard $2 input/$10 output per million token list rates from GPT-6 Sol. Its $0.10 cached-input rate is lower, but long prompts and speed modes can change the bill. Test accepted results and total task cost on your own prompts.
Where can you use GPT-6.1 Sol?
The API model ID is gpt-6.1-sol. OpenAI lists it for the API and eligible Plus, Pro, Business, Enterprise, and Edu access in ChatGPT Work and Codex. That list does not mean the model is selectable in ordinary Chat; the launch says Chat support is coming later. Check your own account, workspace settings, and region before switching.
Context, tools, and reasoning limits
The model documentation lists text and image input, text output, 1,050,000 context tokens, 128,000 maximum output tokens, and a knowledge cutoff of April 30, 2026. Reasoning effort runs from low to max, with medium documented as default; none and minimal are not supported. Use the Responses API for tool calling. Chat Completions does not provide its tool-calling contract.
What does the API cost?
Standard API prices per million tokens are $2 input, $0.10 cached input, $2.50 cache write, and $10 output. Above 272,000 input tokens, OpenAI prices the entire request at $4 input, $0.20 cached input, $5 cache write, and $15 output. Fast mode doubles applicable rates, while Batch and Flex have separate discounts. Check residency and mode availability before assuming the same route works in every region.
How does it compare with GPT-6 Sol?
GPT-6 Sol remains the direct predecessor and the current standard input/output list-rate comparison. GPT-6.1 Sol halves the listed cached-input rate and has a different reasoning-effort range. OpenAI describes performance close to Astra at a fraction of Astra's standard token price, but that vendor comparison does not predict quality or cost on every repository or agent task.
Useful first jobs
- Coding tasks with tests and human review.
- Tool-using agents on the Responses API.
- Long-context analysis with measured cache behavior.
Frequently asked questions
Is GPT-6.1 Sol in ChatGPT Chat?
OpenAI lists GPT-6.1 Sol in the API and eligible ChatGPT Work and Codex plans, and says Chat availability is coming later. Check your account before relying on a rollout.
Does GPT-6.1 Sol cost less than GPT-6 Sol?
Their standard $2 input/$10 output per million token list rates match. GPT-6.1 Sol lists $0.10 cached input versus GPT-6 Sol's $0.20; long-context and mode surcharges can change total task cost.
Sources and evidence
OpenAI's September 29 launch, live API model contract, and pricing page establish identity, limits, prices, and rollout. Near-Astra performance and speed comparisons are OpenAI measurements, not independent guarantees or account-wide access.
- Introducing GPT-6.1 Sol ↗
OpenAI · Primary source
- GPT-6.1 Sol API model ↗
OpenAI · Primary source
- OpenAI API pricing ↗
OpenAI · Primary source