GPT-6.1 Sol

What OpenAI says. The model selection guide suggests Sol "for complex projects where cost matters", giving examples such as creating a board presentation from financial results or building a website from a product brief. It maps Sol at medium effort to complex technical work and coordinated deliverables you expect to revise, and Sol at extra high effort to polished deliverables, connected visual systems and decisions built from conflicting evidence. The Using GPT-6 guide labels Sol the "balanced speed, cost, and intelligence" choice. For latency-sensitive work, the model page points to Ultrafast mode, set with service_tier: "ultrafast" in the Responses API.

Our editorial judgment. Sol looks like the sensible starting point for most teams building on GPT-6: it has the same context window, output limit and tool list as Astra at a fifth of the price, and full US and EU residency support. Start here, then escalate specific task types to Astra only when your tests show Sol failing them. Ultrafast at 6x Standard makes Sol cost more per token than Astra at Standard (USD 12 versus USD 10 input), so reserve it for interactive cases where speed is the product. For coding assistant comparisons, see our AI coding tools guide.

• Long-prompt surcharge. Prompts above 272K input tokens are billed at 2x input and cache rates and 1.5x output for the full request: USD 4 input and USD 15 output per 1M tokens at list price.
• No reasoning-off mode. none and minimal are not supported. The guide says to use low when migrating code that relied on them.
• Tool calling requires Responses. Chat Completions works for Sol only without tool calling.
• Unsupported features. Fine-tuning and predicted outputs are listed as unsupported.
• Unsupported parameters. When reasoning effort is not none, remove temperature, top_p and top_logprobs.
• Data residency. Sol supports US and EU data residency, including with Fast and Ultrafast modes. Regional processing adds a 10% premium where available.
• Rate limits by tier. Build: 5,000 RPM, 1,000,000 TPM. Launch: 10,000 RPM, 4,000,000 TPM. Grow: 15,000 RPM, 40,000,000 TPM. The page lists no batch queue limit for Sol.
• Migration from GPT-6 Sol. The guide notes an earlier gpt-6-sol model and asks existing users to review migration guidance before switching.

All three GPT-6 models share a 1,050,000 token context window, 128,000 max output, text and image input and the same Responses API tool list.

Sol's cached input discount is steeper than its siblings: 5% of input versus 10% for Astra and Luna. That matters for agent loops that resend a long, stable prefix. Luna is twenty times cheaper per token than Sol and has a later knowledge cutoff (18 May 2026). Outside OpenAI, Grok 4.7 lists the same USD 2.00 input price, a lower USD 6.00 output price and a smaller 500,000 token context window, according to its vendor docs.

1. Collect 30 or more representative tasks with known good answers, weighted toward the complex work you would otherwise send to Astra.
2. Run them on Sol at medium effort, then at xhigh, and on Astra at medium, keeping prompts, tools and context identical.
3. Score each output against a written rubric and record tokens, so you can compute cost per accepted result for each setting.
4. If your prompts share a long prefix, measure cache hit rates, since Sol's 5% cached rate can change the cost picture more than the headline price.
5. If you need low latency, run a separate trial with Ultrafast and compare its cost per task against Astra at Standard before committing.

We read the GPT-6.1 Sol model page, the Using GPT-6 guide and the model selection guide on 15 October 2026. We have not run GPT-6.1 Sol, and nothing here is a test result or ranking. Prices and limits change; confirm on the vendor page before budgeting. See our ranking method and the model directory for related profiles.

What is GPT-6.1 Sol? It is OpenAI's mid-tier GPT-6 model, positioned as near-Astra performance at lower cost for complex coding, computer use and professional work.

How much does GPT-6.1 Sol cost? It costs USD 2 per 1M input tokens and USD 10 per 1M output tokens at list price on 15 October 2026, with cached input at USD 0.10 per 1M tokens.

What is the API model name? It is gpt-6.1-sol, which is the only snapshot listed on the model page.

Can I turn reasoning off with GPT-6.1 Sol? No, the none and minimal effort settings are not supported, and medium is the default.

Does GPT-6.1 Sol support EU data residency? Yes, OpenAI lists US and EU data residency for Sol, including with Fast and Ultrafast modes.

Can I use tools through Chat Completions? No, Chat Completions is supported only without tool calling, so tool use requires the Responses API.