OpenAI
OpenAI prices GPT-6.1 Sol at one-fifth of GPT-6 Astra
OpenAI released GPT-6.1 Sol Sept. 29 at $2 input and $10 output per 1M tokens, and claims near-Astra performance without publishing numbers.
OpenAI released GPT-6.1 Sol, a model it says performs near GPT-6 Astra at a fifth of Astra’s listed price. The company published no benchmark numbers for that claim.
The release came Sept. 29, 2026, according to the OpenAI API changelog, read Sept. 30. The API ID is gpt-6.1-sol, and OpenAI describes it as for “complex coding and professional work at a lower cost than GPT-6 Astra.”
Pricing
Standard pricing is $2 per 1M input tokens and $10 per 1M output tokens for prompts up to 272K input tokens, the changelog entry and the GPT-6.1 Sol model page said Sept. 30. Cached input costs $0.10 and cache writes cost $2.50.
The model page lists the modifiers. Prompts over 272K input tokens are billed at 2x the input and cache rates and 1.5x the output rate for the full request. Fast mode costs 2x standard, and Batch and Flex are 50% lower.
OpenAI’s pricing page, read Sept. 30, lists GPT-6 Astra at $10 for input, $1 for cached input, $12.50 for cache writes and $50 for output at standard short-context rates. Sol’s input and output prices are each one-fifth of those.
The Lab Notes reported Sept. 28 that GPT-6 Sol, released Sept. 22, lists the same $2 and $10 prices. Its cached input rate is $0.20, against $0.10 for GPT-6.1 Sol.
What OpenAI says about quality
OpenAI’s only evidence for near-Astra performance is its own description. The model page says Sol delivers “near-Astra performance at a lower cost for complex coding, computer use, and professional work,” and tells developers to compare it with Astra on their own tasks “to assess the tradeoff between quality and cost.” The pages read contained no benchmark numbers.
The Codex changelog entry of Sept. 29, titled “GPT-6.1 Sol in Codex and ChatGPT Work,” repeats the claim. OpenAI said Sol “offers near-Astra performance for complex work at a lower cost than Astra” and suggested it “for repeated, long-running work across code, apps, and documents.” Availability “depends on your plan, client, and workspace settings,” the entry said.
Specs and where it runs
Sol has a 1,050,000-token context window, 128,000 maximum output tokens and an Apr. 30, 2026 knowledge cutoff, according to the model page. It takes text and image input and returns text. Reasoning effort defaults to medium and goes up to max.
The Responses API supports computer use and MCP for the model. The API changelog also lists Multi-agent, in beta, which lets the model delegate work to subagents within a Responses API request.
Codex CLI 0.159.1, also released Sept. 29, made GPT-6.1 Sol the default model in its bundled catalog and in the Amazon Bedrock Mantle and Runtime catalogs, the Codex changelog said.
Analysis
We think the $2 and $10 prices against Astra’s $10 and $50 are large enough that buyers will want to try Sol on their Astra workloads. OpenAI gave no numbers for “near-Astra,” so buyers have to run that comparison themselves, as OpenAI’s own model page invites them to.