OpenAI: GPT-6.1 Sol Pro

GPT-6.1 Sol Pro is a text model from OpenAI, listed on api14 as openai/gpt-6.1-sol-pro and released by the provider on 2026-09-29. It shares its underlying model with GPT-6.1 Sol, but it is served with reasoning.mode set to pro, which aims at higher quality responses on complex tasks. If you have used the standard Sol variant, this is the same model with a different reasoning configuration rather than a separate architecture.

The model accepts text, images and files as input, and it returns text only. Its context window is 1,050,000 tokens, and the maximum output is 128,000 tokens, so long documents and large mixed inputs fit comfortably while still leaving room for a substantial response. Supported request parameters include max_tokens and max_completion_tokens, response_format and structured_outputs, tools and tool_choice, seed, verbosity, and the reasoning controls: reasoning, reasoning_effort and include_reasoning. Note that pro mode spends far more reasoning effort than the standard mode, which is the main trade-off you are making when you pick this variant.

Because the inputs cover text, images and files within a very large context, practical uses lean toward work where the whole problem fits in one request and the answer needs care. Examples include reviewing a full codebase or specification bundle alongside screenshots, extracting structured records from scanned or mixed-format documents, long-form analysis and drafting where earlier material must stay in view, and agent-style workflows that call tools and return structured outputs. Reasoning-heavy tasks such as multi-step debugging, financial or technical document review, and planning work that benefits from deliberate step-by-step thinking are the ones the pro configuration is aimed at.

You call it through api14's OpenAI-compatible API using the model id openai/gpt-6.1-sol-pro, the same way you would call any other chat completions endpoint, so existing OpenAI SDKs and clients work without changes beyond the model name. Pass the parameters listed above in the request body as usual; if you are porting from another OpenAI model, tool definitions, response_format and the token limit fields keep the same shape. Reasoning is configurable, so you can tune reasoning_effort and decide whether to surface the reasoning trace with include_reasoning.

There is no documented restriction on the categories this model supports; it is listed under text only, with no audio or video input and no image output, so a model with those modalities would be a better fit for speech or image generation work. The pro reasoning mode spends far more than the standard mode, so for simple, high-volume or latency-sensitive requests the standard GPT-6.1 Sol is the more natural choice, and this variant is best reserved for the complex tasks it was configured for.