GPT-6.2 Astra is OpenAI's top general model, released alongside GPT-6.2 Sol. OpenAI says it costs significantly less per task than GPT-6 Astra, with input and output prices 20% lower and cached input 60% lower. It reads text and images, holds about 1.05M tokens of context (922K input) and writes up to 128K tokens. Prompts above 272K tokens are billed at a higher long-context rate, and Fast and Ultrafast modes trade more money for lower latency.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| TerminalBench | 64.1 | accuracy | 1d ago |