← The feed Many Minded · The brief
KNOWLEDGE · forward · impact 2/5 · 2026-11-18 · OpenAI

OpenAI's new Astra and Sol models claim stronger results on fewer tokens

OpenAI's GPT-6.2 Astra cuts API prices 20% versus GPT-6 Astra; Sol holds its prior price. Benchmarks are OpenAI's own.

OpenAI has introduced two models: GPT-6.2 Astra, a frontier model aimed at hard projects, and GPT-6.2 Sol, an efficient model for everyday work. The company says both deliver better results on demanding tasks while using fewer output tokens.

On pricing, Astra's API rates are $8 per million input tokens, $0.40 per million cached input tokens and $40 per million output tokens. That is 20% below GPT-6 Astra's input and output prices, with a 60% discount on cached input. Sol keeps the API token prices of GPT-6.1 Sol, so its gains come from performance rather than a price cut.

The cost claims are the substance. On Agents' Last Exam, OpenAI says Astra scores 5.2 points above GPT-6 Astra at under a third of the cost, and 5.5 points above Claude Fable 5.2 at about 5% of the cost. On Cognition's FrontierCode 2.0, Sol lands within 0.5 points of Astra at about one-fifth of the cost. On OpenAI's LifeVal evaluation, Sol beats GPT-6 Astra at about one-seventh of the cost. Customers report savings too: Legora cites 62% fewer tokens than GPT-6.1 Sol on legal-agent tasks, and ClickUp cites about 40% fewer tokens and 30% fewer review rounds.

Why it matters for abundance: if comparable or better capability for work, coding and computer use costs a fraction as much, stronger intelligence reaches people and organizations that could not afford it before. Watch whether independent testing confirms the gaps.

Caveats: the results come from OpenAI's own announcement, and some use internal or OpenAI-run setups. The Fable 5.2 leaderboard result excludes nine tasks that used a fallback model. Both models meet OpenAI's Critical cybersecurity threshold under its Preparedness Framework, so they launch with tighter safeguards that also restrict some legitimate vulnerability research. They are not yet available through Daybreak.

Source: OpenAI