<MCP.Playground/>
online

GPT-6.1 Sol

Runs MCP tools

OpenAI’s upgrade to GPT-6 Sol, released 29 September 2026, a week after the original. It still sits below the flagship Astra, and OpenAI says it comes close to Astra on coding and computer use.

Vendor
OpenAI
Released
29 September 2026
Context
1.05M tokens
Input
Text, image, file

Paste a server URL, pick a model, and watch it call your tools in a real conversation. No install.

No server of your own? Leave it blank and use one of the public mock servers.

MCP Evals

Can GPT-6.1 Sol actually complete tasks with your server?

One chat shows the tool calls work. An eval writes real tasks from your tool schemas, has GPT-6.1 Sol drive each one, and reports where it got the wrong answer even though every call succeeded. Pick GPT-6.1 Sol as the driver.

Run evals →See a sample

What GPT-6.1 Sol is

GPT-6.1 Sol replaces nothing. GPT-6 Sol stays available beside it. OpenAI aims it at agentic coding, computer use, document-heavy professional work and multi-step workflow automation. It has a 1.05M-token window and takes text, image and file input. Sol Pro is the same model served with reasoning mode set to pro. It spends far more reasoning per request and takes much longer to answer.

What OpenAI says

  • DeepSWE v1.1: matches GPT-6 Astra, and 6.4 points above GPT-6 Sol’s best score.
  • AutomationBench: 4.8 points above GPT-6 Sol at medium effort.
  • OSWorld 2.0 (offline set): within 2.1 points of Astra at max effort.

These are the vendor’s own claims, not measurements of ours. Run the model against your own server to find out whether they hold for your tools.

How it reaches MCP

  • The OpenAI Responses API, which can call remote MCP servers as a tool type. Model id gpt-6.1-sol.
  • ChatGPT Work and Codex on the Plus, Pro, Business, Enterprise and Edu plans, where MCP servers connect as connectors
  • MCP Playground’s Agent Studio: paste your server URL above and run Sol or Sol Pro against it in the browser, with no OpenAI key

What to watch for

  • Reasoning effort starts at low. none and minimal are not supported, so a client that sends minimal to keep a tool loop fast has to change that setting before it switches from an older model.
  • Sol Pro spends several times the reasoning of Sol on the same request, by OpenRouter’s description, and takes much longer. Test the one you plan to ship, because the two can choose tools differently on the same prompt.
  • Like GPT-6 Sol, OpenRouter lists seed but not temperature. Use seed if you want repeatable tool calls. Lowering temperature has no effect here.

Run the same multi-step prompt on GPT-6 Sol and GPT-6.1 Sol against the complex-schema mock, for example create a profile and then configure_workflow for it. If 6.1 gets the chain right where 6 Sol did not, that is the upgrade showing on your kind of task. All the mock servers →

Variants available here

ModelModel ID
GPT-6.1 Solopenai/gpt-6.1-sol
GPT-6.1 Sol Proopenai/gpt-6.1-sol-pro

Go deeper

Frequently asked questions

What is the difference between GPT-6.1 Sol and Sol Pro?
Same model, different reasoning mode. Sol Pro runs with reasoning mode set to pro for harder tasks. It spends far more reasoning per request, so runs are slower and use more tokens.
GPT-6.1 Sol or GPT-6 Sol?
OpenAI reports 6.1 ahead on every benchmark it published against 6 Sol, and both stay available. Run the same prompt on both against your server here before you switch, since a better benchmark score does not guarantee the same tool choices.
GPT-6.1 Sol or GPT-6 Astra?
OpenAI says 6.1 Sol matches Astra on DeepSWE and comes within about two points on OSWorld. For most MCP workloads, test Sol first and move up to Astra only if it misses on your tools.
When was GPT-6.1 Sol released?
29 September 2026, one week after GPT-6 Sol and Luna.

Sources

GPT-6.1 Sol for MCP: Specs & Tool Calling — Test Free | MCP Playground