AI model · by OpenAI
GPT-6.1 Sol
GPT-6.1 Sol is OpenAI's model for demanding coding, computer control, and professional tasks, delivering performance close to Astra for less money.
- 1.1M tokens context
- Closed weights
- Checked against OpenAI's page 19 Oct 2026
What GPT-6.1 Sol is for
OpenAI states that this model comes close to Astra's capability while costing less, and it is meant for hard coding, computer operation, and professional tasks. It offers US and EU data residency, covering Fast and Ultrafast modes as well. Ultrafast requests go through the Responses API by setting the model to gpt-6.1-sol and the service tier to ultrafast.
What OpenAI says it does well
- Text and image input with text output.
- Supports streaming, function calling, and structured outputs.
- 1,050,000 token context window with up to 128,000 max output tokens.
Where to use it
- OpenAI API via the Responses API (v1/responses); Chat Completions is supported without tool calling.
- Ultrafast mode in the Responses API with model gpt-6.1-sol and service_tier ultrafast.
- US and EU data residency, including with Fast and Ultrafast modes.
Limits OpenAI mentions
- Fine-tuning is not supported.
- Audio and video modalities are not supported.
- The none and minimal reasoning efforts are not supported.
- Chat Completions is supported without tool calling.