New GPT-6.2 Sol Nears Astra at One-Fifth Cost

OpenAI's new Sol model came within half a point of Astra on one vendor-tested coding benchmark, a result not independently verified.

Updated on Nov. 17, 2026 in Artificial Intelligence

New GPT-6.2 Sol Nears Astra at One-Fifth Cost

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

Why it matters

The release adds a faster output tier and a cheaper high-end model, pushing the cost and speed trade-offs of frontier models toward users who need more output per usage limit. OpenAI says both models can finish demanding tasks with fewer output tokens, which could extend how much work users complete before hitting their limits.

The figures are vendor-reported: Ultrafast mode produces tokens at up to 4.5 times the standard rate, and Sol trailed Astra by half a point on Cognition's FrontierCode 2.0 test at maximum reasoning effort. Independent replication of these numbers has not been reported, and results on other benchmarks were not given.

The players

OpenAI

OpenAI is the San Francisco-based AI lab behind the ChatGPT product line and the GPT model family, offering both consumer subscriptions and developer API access.

The details

GPT-6.2 Astra and Sol are large language models, the systems that generate text one token, or word fragment, at a time. Ultrafast mode is a faster output setting that generates tokens at up to 4.5 times the standard rate, selectable from the model picker. OpenAI says both models can complete demanding tasks with fewer output tokens, meaning less generated text per task, which the company argues lets users get more done before reaching a usage limit.

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.

The Tech Race

The release extends OpenAI's GPT-6 generation, following the September 2026 launch of Astra and the GPT-6.1 upgrade to Sol at DevDay. It places OpenAI's pricing and speed tiers against competing frontier model offerings, though the source does not name specific competitor products.

Both models are available now on Plus, Pro, Enterprise and Edu plans through ChatGPT Work and Codex, and through the API for developers. The $200 monthly Pro plan gets Ultrafast mode with Sol for the first time, and the $500 Pro plan gets Ultrafast mode on both models.

The takeaway

The pattern to watch is whether a cheaper high-end model keeps close benchmark parity, which Sol's half-point gap to Astra on FrontierCode 2.0 suggests. Readers can check whether independent evaluations of FrontierCode 2.0 confirm the vendor-reported figures in the weeks after release.

Further reading

For more on the models behind this release, see the Artificial Intelligence section.