OpenAI Releases GPT-6.2 Astra and Sol Models

Both models are now available in ChatGPT Work, Codex, and through the API, with faster Ultrafast modes on higher-tier plans.

Updated on Nov. 17, 2026 in Artificial Intelligence

OpenAI Releases GPT-6.2 Astra and Sol Models

OpenAI released its GPT-6.2 Astra and GPT-6.2 Sol models in ChatGPT Work and Codex on November 17, 2026, and made them available to developers through the API. Both models are offered on Plus, Pro, Enterprise, and Edu plans.

Why it matters

OpenAI says both models can complete demanding tasks with fewer output tokens, letting users accomplish more before reaching their usage limits. The release also changes the cost comparison between the two models, with Astra priced 20% lower than GPT-6 Astra and Sol keeping its GPT-6.1 pricing.

OpenAI's claims rest on its own product announcement, with Ultrafast mode generating tokens at up to 4.5 times the standard rate. On Cognition's FrontierCode 2.0 test at maximum reasoning effort, Sol came within half a point of Astra, though the source does not report results from independent evaluators.

The players

OpenAI

OpenAI is an artificial intelligence research and deployment company that develops the GPT family of language models and the ChatGPT product.

The details

Ultrafast mode is available on the $200-per-month Pro plan with Sol and on the $500-per-month Pro plan with both models, and users can switch speeds from the model picker. OpenAI's announcement frames the models as tools for finishing demanding tasks with fewer output tokens.

Timeline

  1. September 2026: OpenAI launched the GPT-6 generation with Astra.

  2. A few weeks after September 2026, at DevDay: OpenAI upgraded Sol to GPT-6.1.

  3. November 17, 2026: GPT-6.2 Astra and Sol were released in ChatGPT Work, Codex, and the API.

Roadmap

The GPT-6.2 release extends the GPT-6 generation that OpenAI launched in September 2026 with Astra, following the GPT-6.1 upgrade to Sol at DevDay. The new versions revise pricing, with Astra now 20% cheaper than its GPT-6 predecessor while Sol keeps its GPT-6.1 rate.

Users on Pro plans can switch to Ultrafast mode from the model picker to generate tokens faster, with the $200 plan limited to Sol and the $500 plan covering both models. Developers can access Astra and Sol through the API today, with Astra's lower prices making it a cheaper option than its predecessor.

The takeaway

Choosing between Astra and Sol involves a tradeoff in cost and capability: Sol costs about one-fifth as much as Astra and came within half a point of it on one coding benchmark at maximum reasoning effort. Users weighing price against performance should test both on their own workloads.

Further reading

For broader context on how these models fit into the field, explore our coverage of Artificial Intelligence.