OpenAI's GPT-6.2 Astra ranks second on Artificial Analysis index at 62, behind Claude Fable 5.2 (65)
OpenAI released GPT-6.2 Astra and GPT-6.2 Sol. Artificial Analysis ranks Astra second in its Intelligence Index with 62 points, behind Claude Fable 5.2 at 65, and finds coding performance flat versus the previous generation. Astra's API price is $8 per 1M input tokens and $40 per 1M output tokens.
OpenAI's GPT-6.2 Astra scores 62 on the Artificial Analysis Intelligence Index, second only to Anthropic's Claude Fable 5.2 at 65. OpenAI released Astra and a second model, GPT-6.2 Sol, on November 17, 2026, as reported by The Decoder. Artificial Analysis also finds no coding gain for either model. Epoch AI has Astra tied with Fable 5.2, according to The Decoder.
Availability and pricing
Both models are available to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and through the OpenAI API. Pro-500 subscribers get Ultrafast mode with both models. Pro-200 subscribers get Ultrafast access for the first time, starting with GPT-6.2 Sol.
| Model | Input (per 1M tokens) | Cached input | Output (per 1M tokens) |
|---|---|---|---|
| GPT-6.2 Astra | $8 | $0.40 | $40 |
| GPT-6.2 Sol | $2 | $0.10 | $10 |
OpenAI says Astra's API pricing is 20 percent lower than its predecessor's. Sol keeps GPT-6.1 Sol's pricing. Context window and parameter count are not disclosed in the reporting.
Positioning
OpenAI describes Astra as built for demanding professional work in science, law and finance, and calls it "the world's best model for scientists." Sol is aimed at fast, everyday tasks. OpenAI says Astra was trained to handle professional conventions in law, finance and accounting more reliably, including slides that follow template layouts, colors and typography.
OpenAI's benchmark claims
The following figures come from OpenAI:
- Agents' Last Exam: Astra scores 5.2 points above GPT-6 Astra at less than a third of the cost. It leads Claude Fable 5.2 by 5.5 points at 5 percent of the cost. Only Google's Gemini 4 Carbon keeps up.
- Harvey LAB-AA (legal): Astra more than doubles Claude Fable 5.2's pass rate at about one-sixth of the cost.
- GDPval-v2 (internal): Astra gains 6.2 percentage points over its predecessor with 22 percent fewer output tokens.
- AA-3D: Astra leads with 1,318 Elo, ahead of Claude Fable 5.2 (1,271) and GPT-6.2 Sol (1,259).
- Terminal-Bench-Science: Astra is 9 percentage points above GPT-6 Astra at less than a third of the cost.
- FrontierCode 2.0: Sol trails Astra by 0.5 points at one-fifth the cost.
Independent results
In the Artificial Analysis Intelligence Index, which aggregates nine evaluations, Astra gains five points over GPT-6 Astra. It passes Gemini 4 Carbon (60) and Claude Opus 5.5 (59). Sol rises two points to 57, matching GPT-6 Astra at about 15 percent of the cost per task, according to Artificial Analysis.
Artificial Analysis says OpenAI models dominate the cost-versus-performance Pareto frontier, with only Anthropic and Meta models also on it. On efficiency, measured in output tokens per task, Claude Fable 5.2 is the only outside model among the leaders.
Astra leads on Terminal-Bench-Science 0.1 and on AA-Robotics v0.1, where it scores 79 against Sol's 78 and Claude Fable 5.2's 75. It places second behind Claude Fable 5.2 on CritPt v2, AA-Omniscience and HLE-Diamond.
Coding is flat. In the Coding Agent Index, run with the Codex harness at max compute, Astra stays level with GPT-6 Astra. Sol drops three points from GPT-6.1 Sol, driven by a 14-point fall on SWE-Atlas-QnA-AA, where it omits key facts from answers more often. Artificial Analysis also flags a high hallucination rate for Astra at maximum effort.
Cybersecurity
Both models reach the "Critical" threshold for cybersecurity in OpenAI's Preparedness Framework, as GPT-6 Astra did. OpenAI is adding stricter safeguards for risky cybersecurity requests and acknowledges these will also limit legitimate vulnerability research and security testing. On the internal ExploitBench test, Astra clears a 50 percent success rate, far above GPT-6 Astra, with safeguards and classifiers turned off. OpenAI plans to offer the models through OpenAI Daybreak, which gives vetted security researchers access to models with fewer guardrails.
What this means
The independent picture is more measured than OpenAI's framing. Artificial Analysis confirms OpenAI's science and factual-knowledge claims and places Astra second overall at 62, with Sol at 57 for a fraction of the cost. The gains are concentrated in science, 3D design, robotics and cost efficiency. Coding is unchanged for Astra and slightly lower for Sol, so teams choosing a model mainly for coding agents should not assume improvement from the version bump. The high hallucination rate Artificial Analysis flags at maximum effort is a point to test in factual, high-stakes workflows. For API buyers, the practical change is Astra's 20 percent lower price. Security teams should note that stricter safeguards apply by default, with Daybreak as the route for vetted researchers.
Related Articles
OpenAI's GPT-6.2 Astra and GPT-6.2 Sol are generally available in GitHub Copilot
OpenAI's GPT-6.2 Astra and GPT-6.2 Sol are now generally available in GitHub Copilot, according to the GitHub changelog. GitHub says both models are designed for agentic coding and code review, with stronger coding judgment than earlier GPT-6 models.
OpenAI Reports Up to 14% PPA Gain on AMD Flagship GPU IP From Model-Rewritten RTL
OpenAI reported on November 18, 2026 that its models achieved up to a 14% improvement in power, performance and area (PPA) on AMD flagship GPU intellectual property, measured against heavily optimized baselines. Some of the circuit rewrites were included in a design sent for manufacturing.
OpenAI models optimized circuits used in Jalapeño silicon and improved RISC-V core designs
OpenAI reports that its models optimized circuits used in Jalapeño silicon and improved RISC-V core designs. The company presents the work as progress in AI-assisted chip design with GPT-6 Astra.
OpenAI Python SDK v3.53.0 adds GPT-6.2 Astra, Astra Law and Sol identifiers
OpenAI Python SDK v3.53.0, dated 2026-11-17, adds model identifiers for GPT-6.2 Astra, GPT-6.2 Astra Law and GPT-6.2 Sol. It also adds 2-hour and 8-hour Responses background windows.
Comments
Loading...