OpenAI: GPT-6.1 Sol

GPT-6.1 Sol is a text model from OpenAI, released by the provider on 29 September 2026. It is an upgrade to GPT-6 Sol and sits below the flagship GPT-6 Astra within the GPT-6 series. The provider positions it for agentic coding, computer use, and document-heavy professional work, which places it in the mid tier of the current GPT-6 lineup rather than at the top.

The model accepts text, images, and files as input and returns text as output. Its context window is 1,050,000 tokens, and the maximum output length is 128,000 tokens. That combination means you can submit very large inputs, such as long documents or mixed text and image payloads, and still receive a substantial text response in a single call. Supported request parameters include include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, tool_choice, tools, and verbosity.

Those capabilities point to a few practical uses. Agentic coding fits naturally, since the model supports tools and tool_choice, letting it call functions and act on results across multiple steps. Computer use scenarios follow from the same tool support combined with image input, for example interpreting screenshots or interface state and deciding on the next action. Document-heavy professional work suits the million-token context and file input, so you can pass long reports, contracts, or reference sets without chunking. Structured outputs, response_format, and seed help when you need predictable, machine-readable results or reproducible runs, and reasoning_effort and verbosity let you tune how much internal reasoning the model does and how much it says.

You call the model through api14 using the model id "openai/gpt-6.1-sol". The API is OpenAI-compatible, so existing OpenAI client libraries and request shapes work with only the base URL and API key pointed at api14. In the request body you set the model field to that id and can use any of the parameters listed above, such as tools and tool_choice for function calling, response_format or structured_outputs for structured replies, and max_completion_tokens or max_tokens to cap output. Because the page shows a live price table, check there for current rates rather than relying on any fixed figure.

One thing to weigh is output length: responses are capped at 128,000 tokens, so tasks needing a single very long generation may need to be split. It is also positioned below GPT-6 Astra, so workloads where Astra's extra capability matters may be better served there. Apart from that, the wide context, multimodal input, and tool support make it a reasonable default for agentic and document-heavy text work.