Quick answer
Use Astra when the task needs OpenAI's highest capability or very long context; start with Sol for demanding everyday coding when cost matters.
- Released
- September 3, 2026
- API model
- gpt-6-astra
- Price
- $10 in / $50 out per MTok; $1 cached input; long-context rates apply above 272K input tokens
- Context
- 1,050,000 tokens (128K max output)
01
GPT-6 Astra: verified facts
| Field | Verified value |
|---|---|
| Vendor | OpenAI |
| Release | September 3, 2026 |
| API identifier | gpt-6-astra |
| Input / output price | $10 in / $50 out per MTok |
| Cache / fast pricing | $1 cached input; long-context rates apply above 272K input tokens |
| Context / max output | 1,050,000 tokens (128K max output) |
| Availability | Highest-capability GPT-6 option where available |
Every number on this page was checked against the linked vendor documentation on September 23, 2026. Prices are API list prices per million tokens; product subscriptions and regional availability are separate.
02
Where this model fits
OpenAI positions Astra above Sol and Luna for maximum capability. It accepts a 1.05M-token context, but prompts above 272K tokens receive long-context pricing. Astra does not support the none reasoning setting.
Your AI workspace for shipping software.
Download AIDEN free and point it at your existing Claude Code or Codex setup. No credit card, running in minutes.
Download AIDEN freeFree to start · macOS 12+ · No credit card required
03
When to use it
Use case 1
Ambiguous end-to-end engineering work
Use case 2
Large migrations spanning code, browsers, and documents
Use case 3
The hardest tasks where quality matters more than token price
| Strengths | Trade-offs |
|---|---|
| OpenAI's highest-capability GPT-6 model | Costs five times more than GPT-6 Sol per input and output token |
| 1.05M-token context and 128K maximum output | Does not support the none reasoning effort |
| Full Responses API tool support including computer use, shell, MCP, and skills | Requests above 272K input tokens use higher long-context rates |
04
Published benchmarks
| Benchmark | Score | Source / caveat |
|---|---|---|
| Terminal-Bench 4.0 | 57.9% | Anthropic comparison table, score reported by OpenAI |
| FrontierCode v1.1 | 53.3% | Anthropic comparison table, score reported by OpenAI |
05