<MCP.Playground/>
online

GPT-5.6 Sol

Runs MCP tools

The flagship of OpenAI’s GPT-5.6 series, generally available from 9 July 2026. GPT-6 has since replaced it at the top of the line-up, but plenty of agents still run on it, so it is the model to compare against before you upgrade.

Vendor
OpenAI
Released
9 July 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-5.6 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-5.6 Sol drive each one, and reports where it got the wrong answer even though every call succeeded. Pick GPT-5.6 Sol as the driver.

Run evals →See a sample

What GPT-5.6 Sol is

GPT-5.6 Sol is the top tier of GPT-5.6, which shipped as Sol, Terra and Luna after a preview that began on 26 June 2026. OpenAI pitches it at complex reasoning, coding and agent work, and in particular at command-line and multi-step coding tasks. 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, which spends far more reasoning on each answer.

What OpenAI says

  • Terminal-Bench 2.1: 88.8%. Terra scores 84.3% and Luna 82.5%.
  • ExploitBench: competitive with Mythos Preview while using about a third of the output tokens.
  • GeneBench v1: ahead of GPT-5.5 while using fewer tokens.

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
  • ChatGPT, where MCP servers connect as connectors on paid plans
  • 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

  • OpenRouter lists seed but neither temperature nor stop for GPT-5.6 Sol, the same as for GPT-6. A client that lowers temperature to make tool calls repeatable gets no effect. Use seed instead.
  • Sol Pro is not a better model, it is a longer think. It costs more time and output per turn, so check that it changes which tools get called before you pay for it.
  • If you are moving to GPT-6 Sol or GPT-6.1 Sol, the request parameters are the same, so the client needs no change. Differences in the run come from the model.

Run the same four-step task on GPT-5.6 Sol and GPT-6.1 Sol against the complex-schema mock: create_user_profile, process_order, analyze_data, then configure_workflow. If 5.6 Sol already gets every step right, the upgrade is about cost and speed, not whether your server works. All the mock servers →

Variants available here

ModelModel ID
GPT-5.6 Solopenai/gpt-5.6-sol
GPT-5.6 Sol Proopenai/gpt-5.6-sol-pro

Go deeper

Frequently asked questions

What is the difference between GPT-5.6 Sol and Sol Pro?
Same model, different reasoning mode. Sol Pro runs with reasoning mode set to pro, so it reasons for longer on each answer. Runs are slower and use far more tokens.
GPT-5.6 Sol or GPT-6 Sol?
GPT-6 Sol is the newer model, released 22 September 2026, and GPT-6.1 Sol followed it. Both take the same request parameters as 5.6 Sol, so you can compare them on your own server here without changing anything but the model.
Does GPT-5.6 Sol work with MCP servers?
Yes. OpenAI’s Responses API can call remote MCP servers, ChatGPT supports them as connectors on paid plans, and any MCP client can drive it through function calling.
What are GPT-5.6 Terra and Luna?
The other two tiers of GPT-5.6. Terra is the balanced tier, aimed at GPT-5.5-class performance for less. Luna is the fast, low-cost tier.

Sources

GPT-5.6 Sol for MCP: Specs & Tool Calling — Test Free | MCP Playground