GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.
Specification
openai listing
Context
1.1M
Max output
128K
Released
2026-09-04
Knowledge cutoff
2026-04-30
Retires
—
Open weights
no
Input
text, image, pdf
Output
text
Price
US dollars per million tokens · openai list price
Input
$10.00
Output
$50.00
Cache read
$1.00
Cache write
$12.50
Batch and priority prices
US dollars per million tokens · per listing
| Price | Listing | $/M |
|---|---|---|
| Batch input | openai | $5.00 |
| Batch output | openai | $25.00 |
| Priority input | openai | $20.00 |
| Priority output | openai | $100.00 |
Batch input is 50% below the interactive input price at the same listing.
Long-context price tiers
the rate above each context threshold
| Above | Price | $/M |
|---|---|---|
| 272K | Cache read | $2.00 |
| 272K | Cache write | $25.00 |
| 272K | Input | $20.00 |
| 272K | Output | $75.00 |
Quality
11 benchmarks · 2 sources
Composite
100th
percentile of 334 scored models
Rank
2 / 334
±0.07 sd
Evidence
11 × 2
3 or more benchmarks, from 2 or more sources
Effort range
9.8
points between effort settings
coding65th1/3 bench
reasoning100th4/4 bench
math99th3/7 bench
knowledge100th1/1 bench
general96th1/1 bench
By reasoning effort
every setting placed on the same scale as the leaderboard
| Setting | Composite | Benchmarks behind it |
|---|---|---|
| max | 100th | 9 |
| xhigh | 100th | 1 |
| medium | 100th | 1 |
| low | 99th | 1 |
| high | 97th | 2 |
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| Mystery Game Puzzles | 84.0 ±3.7 | max | Epoch AI | 2026-08-30 |
| Chess Puzzles | 72.0 ±4.5 | max | Epoch AI | 2026-08-30 |
| EBR-bench | 76.2 ±8.4 | max | Epoch AI | 2026-08-30 |
| FrontierMath-Tier-4-v2 | 97.6 ±2.4 | high | Epoch AI | 2026-08-30 |
| ↳ | 97.6 ±2.4 | max | Epoch AI | 2026-08-30 |
| ↳ | 97.6 ±2.4 | xhigh | Epoch AI | 2026-08-30 |
| ↳ | 97.6 ±2.4 | medium | Epoch AI | 2026-08-30 |
| ↳ | 87.8 ±5.2 | low | Epoch AI | 2026-08-30 |
| Settings differ by 9.8 points on this benchmark. | ||||
| FrontierMath-Erdos | 2.9 ±2.1 | max | Epoch AI | 2026-08-28 |
| SimpleQA Verified | 75.6 ±1.4 | max | Epoch AI | 2026-08-30 |
| FrontierMath-Tiers-1-3-v2 | 93.7 ±1.4 | max | Epoch AI | 2026-08-30 |
| LiveBench | 83.0 | max · 23 runs | LiveBench | 2026-06-25 |
| GPQA diamond | 95.8 ±1.4 | max | Epoch AI | 2026-08-30 |
| OTIS Mock AIME 2024-2025 | 100.0 ±0.0 | max | Epoch AI | 2026-08-30 |
| MirrorCode | 46.7 | high | Epoch AI | 2026-08-30 |
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 27 hosts
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| openaivendor | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M 922K | 128K | — |
| 302ai | $10.00 | $50.00 | — | — | 1.1M | 128K | — |
| amazon-bedrock · global | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| azure | $10.00 $11.00 | $50.00 $55.00 | $1.00 $1.10 | $12.50 $13.75 | 1.1M 922K | 128K | — |
| azure_ai | $10.00 | $50.00 | $1.00 | $12.50 | 922K | 128K | — |
| bedrock_converse · global | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| crossmodel | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| edenai | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| github-copilot | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| kilo | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| llmgateway | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 1.1M | — |
| llmgateway-providers | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| llmgateway-providers · aws-mantle | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| merge-gateway | $10.00 | $50.00 | $1.00 | — | 1.1M | 128K | — |
| nano-gpt | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| neon | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| ofox | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| opencode | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| openrouter | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| requesty | $10.00 | $50.00 | $1.00 | — | 1.1M | 128K | — |
| vercel | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| vivgrid | $10.00 | $50.00 | $1.00 | $12.50 | 1.1M | 128K | — |
| amazon-bedrock | $11.00 | $55.00 | $1.10 | $13.75 | 1.1M | 128K | — |
| amazon-bedrock · us | $11.00 | $55.00 | $1.10 | $13.75 | 1.1M | 128K | — |
| bedrock_converse · us | $11.00 | $55.00 | $1.10 | $13.75 | 1.1M | 128K | — |
| bedrock_mantle · bedrock-mantle | $11.00 | $55.00 | $1.10 | $13.75 | 1.1M | 128K | — |
| infer · official | $12.50 | $62.50 | $1.25 | $15.625 | 272K | 128K | — |
Price history
append-only observations · a listing writes a row only when its price moves
| Date | Host | Field | $/M |
|---|---|---|---|
| 2026-09-04 | openai | Cache read | $1.00 |
| 2026-09-05 | openai | Cache read | $1.00 |
| 2026-09-04 | openai | Cache write | $12.50 |
| 2026-09-05 | openai | Cache write | $12.50 |
| 2026-09-04 | openai | Input | $10.00 |
| 2026-09-05 | openai | Input | $10.00 |
| 2026-09-04 | openai | Output | $50.00 |
| 2026-09-05 | openai | Output | $50.00 |
| 2026-09-05 | openrouter | Cache read | $1.00 |
| 2026-09-06 | openrouter | Cache read | $1.00 |
| 2026-09-05 | openrouter | Cache write | $12.50 |
| 2026-09-06 | openrouter | Cache write | $12.50 |
| 2026-09-05 | openrouter | Input | $10.00 |
| 2026-09-06 | openrouter | Input | $10.00 |
| 2026-09-05 | openrouter | Output | $50.00 |
| 2026-09-06 | openrouter | Output | $50.00 |
| 2026-09-06 | azure | Cache read | $1.00 |
| 2026-09-07 | azure | Cache read | $1.00 |
| 2026-09-06 | azure | Cache write | $12.50 |
| 2026-09-07 | azure | Cache write | $12.50 |
| 2026-09-06 | azure | Input | $10.00 |
| 2026-09-07 | azure | Input | $10.00 |
| 2026-09-06 | azure | Output | $50.00 |
| 2026-09-07 | azure | Output | $50.00 |
More from OpenAI
most-hosted first