GPT-5.6 Terra
- Context
- 1.05M
- 128K max output
- Input price
- $2.00
- per 1M tokens · OpenAI API
- Output price
- $12.00
- per 1M tokens
- Epoch Capabilities Index
- 159.3
- #7 of 71 listed
- LMArena Text
- 1,465
- #30 of 79 listed
Where to run it
3 offerings · prices as listed| Meter | Tier | Price |
|---|---|---|
| Input | Standard | $2.00 / 1M tokens |
| Batch | $1.00 / 1M tokens | |
| Flex tier | $1.00 / 1M tokens | |
| Fast tier | $4.00 / 1M tokens | |
| Output | Standard | $12.00 / 1M tokens |
| Batch | $6.00 / 1M tokens | |
| Flex tier | $6.00 / 1M tokens | |
| Fast tier | $24.00 / 1M tokens | |
| Cache read | Standard | $0.20 / 1M tokens |
| Batch | $0.10 / 1M tokens | |
| Flex tier | $0.10 / 1M tokens | |
| Fast tier | $0.40 / 1M tokens | |
| Cache write | Standard | $2.50 / 1M tokens |
| Batch | $1.25 / 1M tokens | |
| Flex tier | $1.25 / 1M tokens | |
| Fast tier | $5.00 / 1M tokens | |
Azure OpenAI in Foundry (opens in a new tab) gpt-5-6-terra | ||
| Input | Standard | $2.00 / 1M tokens |
| Priority tier | $4.00 / 1M tokens | |
| Output | Standard | $12.00 / 1M tokens |
| Priority tier | $24.00 / 1M tokens | |
| Cache read | Standard | $0.20 / 1M tokens |
| Priority tier | $0.40 / 1M tokens | |
| Cache write | Standard | $2.50 / 1M tokens |
| Priority tier | $5.00 / 1M tokens | |
Azure OpenAI in Foundry (opens in a new tab) gpt-5-6-terra | ||
| Input | Priority tier | $4.40 / 1M tokens |
| Output | Priority tier | $26.40 / 1M tokens |
| Cache read | Priority tier | $0.44 / 1M tokens |
| Cache write | Priority tier | $5.50 / 1M tokens |
Changes
All changes →- 4 prices dropped on OpenAI API Input (Prompts over 272K tokens, Fast tier) $8.00 · Cache read (Prompts over 272K tokens, Fast tier) $0.80 · Cache write (Prompts over 272K tokens, Fast tier) $10.00 · Output (Prompts over 272K tokens, Fast tier) $36.00
Vendor-reported scores
Published by OpenAI, under the conditions it states| Benchmark | Setting | Score |
|---|---|---|
| SWE-bench Pro | — | 63.4% |
| DeepSWE 1.1 | — | 69.6% |
| Terminal-Bench 2.1 | single agent | 87.4% |
| OSWorld 2.0 | set and scoring not stated | 50.2% |
| BrowseComp | single agent | 87.5% |
| GPQA Diamond | — | 92.9% |
| MMMU-Pro | no tools | 80.7% |
| with tools | 82% | |
| MRCR v2 | 8-needle, 256K–512K | 89.6% |
| 8-needle, 512K–1M | 72.5% |
Cited scores
From third-party leaderboards, not measured by us- Epoch Capabilities Index 159.3
#7 of 71 listed models · CI 156.7–162.3
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 1,465
#30 of 79 listed models · xhigh effort · CI 1,461–1,470 · 35,323 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 86%
#6 of 56 listed models · max effort · ±2.06
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 93.3%
#11 of 64 listed models · max effort · ±1.54
Also listed: low effort 87.4% · reasoning off 77.3%
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
#27 of 79 listed models · xhigh effort · CI 1,504–1,527 · 2,786 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 1,516
#29 of 79 listed models · xhigh effort · CI 1,509–1,523 · 9,671 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made
- LMArena Vision 1,270
#25 of 49 listed models · xhigh effort · CI 1,262–1,278 · 7,726 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 13, 2026 · changes we made