GPT-6 Astra
- Context
- 1.05M
- 128K max output
- Input price
- $10.00
- per 1M tokens · OpenAI API
- Output price
- $50.00
- per 1M tokens
- Epoch Capabilities Index
- 166.6
- #1 of 71 listed
- LMArena Text
- 1,478
- #20 of 79 listed
Where to run it
3 offerings · prices as listed| Meter | Tier | Price |
|---|---|---|
| Input | Standard | $10.00 / 1M tokens |
| Batch | $5.00 / 1M tokens | |
| Flex tier | $5.00 / 1M tokens | |
| Fast tier | $20.00 / 1M tokens | |
| Output | Standard | $50.00 / 1M tokens |
| Batch | $25.00 / 1M tokens | |
| Flex tier | $25.00 / 1M tokens | |
| Fast tier | $100.00 / 1M tokens | |
| Cache read | Standard | $1.00 / 1M tokens |
| Batch | $0.50 / 1M tokens | |
| Flex tier | $0.50 / 1M tokens | |
| Fast tier | $2.00 / 1M tokens | |
| Cache write | Standard | $12.50 / 1M tokens |
| Batch | $6.25 / 1M tokens | |
| Flex tier | $6.25 / 1M tokens | |
| Fast tier | $25.00 / 1M tokens | |
Azure OpenAI in Foundry (opens in a new tab) gpt-6-astra | ||
| Input | Standard | $10.00 / 1M tokens |
| Output | Standard | $50.00 / 1M tokens |
| Cache read | Standard | $1.00 / 1M tokens |
| Cache write | Standard | $12.50 / 1M tokens |
Azure OpenAI in Foundry (opens in a new tab) gpt-6-astra | ||
| Input | Standard | $11.00 / 1M tokens |
| Output | Standard | $55.00 / 1M tokens |
| Cache read | Standard | $1.10 / 1M tokens |
| Cache write | Standard | $13.75 / 1M tokens |
Changes
All changes →- 16 new prices on OpenAI API Input (Batch) $5.00 · Cache read (Batch) $0.50 · Cache write (Batch) $6.25 · Output (Batch) $25.00 · +12 more
- Now priced on Azure OpenAI in Foundry, data-zone Input $11.00 · Cache read $1.10 · Cache write $13.75 · Output $55.00
- Now priced on Azure OpenAI in Foundry Input $10.00 · Cache read $1.00 · Cache write $12.50 · Output $50.00
Vendor-reported scores
Published by OpenAI, under the conditions it states| Benchmark | Setting | Score |
|---|---|---|
| Terminal-Bench 4.0 | maximum across effort levels | 57.9% |
| DeepSWE 1.1 | maximum across effort levels | 74.1% |
| OSWorld 2.0 | v2026.08.08 offline set, partial score; maximum across effort levels | 72.6% |
| BrowseComp | maximum across effort levels | 91.5% |
| GPQA Diamond | maximum across effort levels | 96% |
| Humanity's Last Exam | with tools; maximum across effort levels | 57.2% |
| ARC-AGI-2 | maximum across effort levels | 95% |
| MRCR v2 | 8-needle, 256K–512K | 100% |
| 8-needle, 512K–1M | 96.3% |
Cited scores
From third-party leaderboards, not measured by us- Epoch Capabilities Index 166.6
#1 of 71 listed models · CI 163–172
Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Sep 26, 2026 · changes we made
- LMArena Text 1,478
#20 of 79 listed models · max effort · CI 1,470–1,485 · 7,185 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made
- Epoch AI · FrontierMath Tiers 1–3 93.7%
#1 of 56 listed models · max effort · ±1.44
Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Sep 26, 2026 · changes we made
- Epoch AI · GPQA Diamond 95.8%
#1 of 64 listed models · max effort · ±1.37
Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Sep 26, 2026 · changes we made
- LMArena Text · Chinese 1,516
#25 of 79 listed models · max effort · CI 1,492–1,541 · 531 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made
- LMArena Text · Coding 1,542
#5 of 79 listed models · max effort · CI 1,528–1,557 · 1,738 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made
- LMArena Vision 1,279
#21 of 49 listed models · max effort · CI 1,262–1,296 · 1,367 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 13, 2026 · changes we made