OpenAIGPT-5.6 Terra
Best configuration ranks #35 of 610 on the BenchLeader Index at 64.2 ±3.8 (xhigh reasoning effort). Last measured 8 Sept 2026. Released 9 Jul 2026.
- Blended price
- $4.50/M
- $2.00 in · $12.00 out
- Output speed
- 93 tok/s
- First answer
- 33 s
- first token 2.75 s
- Context
- 1M
- Overall index64
- Reasoning58
- Coding64
- Agents & tools70
- Maths71
- Knowledge61
- Instruction following65
- Human preference66
- Multimodal63
- Long context66
- Composite77
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 52.1 | #257 | 83 tok/s | 0.95 s | $0.0044 | Coding 50 · Composite 57 · Knowledge 53 · Long context 55 · Maths 47 · Multimodal 51 · Reasoning 57 |
| low | 58.8 | #116 | 83 tok/s | 1.95 s | $0.0044 | Agents & tools 72 · Coding 57 · Composite 64 · Instruction following 59 · Knowledge 61 · Long context 62 · Maths 63 · Multimodal 61 · Reasoning 50 |
| medium | 58.1 | #129 | 80 tok/s | 2.07 s | $0.0044 | Coding 57 · Composite 70 · Instruction following 62 · Knowledge 62 · Long context 63 · Multimodal 61 · Reasoning 50 |
| high | 63.1 | #49 | 83 tok/s | 3.08 s | $0.0044 | Agents & tools 84 · Coding 64 · Composite 72 · Instruction following 63 · Knowledge 62 · Long context 65 · Multimodal 64 · Reasoning 59 |
| xhighbest | 64.2 | #35 | 93 tok/s | 33 s | $0.0044 | Agents & tools 70 · Coding 64 · Composite 77 · Human preference 66 · Instruction following 65 · Knowledge 61 · Long context 66 · Maths 71 · Multimodal 63 · Reasoning 58 |
| max | 58.0 | #133 | 97 tok/s | 196 s | $0.0048 | Agents & tools 45 · Coding 63 · Composite 60 · Knowledge 55 · Maths 72 · Reasoning 64 |
| default | 62.3 | #58 | 97 tok/s | 196 s | $0.0044 | Agents & tools 84 · Coding 57 · Composite 82 · Human preference 55 · Instruction following 69 · Knowledge 64 · Long context 68 · Multimodal 65 · Reasoning 49 |
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | 77.3%#124 | 87.4%#61 | – | – | – | 93.3%#14 | – | Epoch AI Benchmarking Hub | |
| SimpleBench | – | – | – | – | 48.9%#45 | – | 48.9%#45 | SimpleBench | |
| LMArena Hard Prompts | – | – | – | – | 1491#40 | – | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | – | – | – | 90.6%#6 | – | LiveBench | |
| GPQA Diamond (AA)not in index | 74.7%#249 | 84.3%#142 | 87.2%#101 | 89.6%#71 | 90.8%#55 | – | 92.5%#35 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 11.4%#230 | 29.2%#111 | 33.3%#94 | 38.5%#64 | 41.9%#46 | – | 42.9%#34 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | 90.9%#24 | – | – | Vals AI | |
| Kagi LLM Benchmark | – | – | – | – | – | – | 51.3%#82 | Kagi LLM Benchmark | |
| ARC-AGI-1 | – | 60.2%#108 | 77.0%#87 | 92.0%#40 | 94.0%#29 | 96.5%#13 | – | ARC Prize | |
| ARC-AGI-2 | – | 18.8%#95 | 37.5%#82 | 67.1%#43 | 74.2%#31 | 83.9%#23 | – | ARC Prize | |
| ARC-AGI-3 | – | 0.0%#38 | 0.1%#35 | 0.5%#22 | 0.7%#21 | 0.8%#20 | – | ARC Prize |
Coding
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SciCode | 44.6%#89 | 49.2%#65 | 49.6%#64 | 50.1%#59 | 51.6%#49 | 53.9%#36 | – | SciCode | |
| WeirdML | – | – | – | 78.3%#17 | – | – | – | WeirdML | |
| FrontierCode | – | – | – | – | – | – | 41.3%#14 | Cognition | |
| LMArena Coding | – | – | – | – | 1519#31 | – | – | LMArena | |
| LMArena WebDev | – | – | – | – | 1521#33 | – | – | LMArena | |
| LiveBench Codingnot in index | – | – | – | – | – | 78.3%#21 | – | LiveBench | |
| SciCode (AA)not in index | – | 49.9%#71 | 50.5%#65 | 52.4%#53 | 52.3%#54 | – | 55.0%#34 | Artificial Analysis | |
| LiveCodeBench | – | – | – | – | 85.9%#25 | – | – | Vals AI | |
| IOI | – | – | – | – | 65.3%#7 | – | – | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | – | 95.4%#5 | – | Vals AI |
Agents & tools
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | – | – | – | 21.5%#54 | – | Terminal-Bench | |
| LMArena Agent | – | – | – | – | 1.5#22 | – | – | LMArena | |
| LiveBench Agentic Codingnot in index | – | – | – | – | – | 55.0%#22 | – | LiveBench | |
| Terminal-Bench Hard | – | 43.9%#35 | – | 57.6%#11 | 62.9%#2 | – | 57.6%#11 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | – | 60.5%#173 | 72.8%#142 | 78.4%#126 | 80.4%#119 | – | 86.3%#87 | Artificial Analysis | |
| Terminal-Bench 2.1 (Vals) | – | – | – | – | – | 77.5%#11 | – | Vals AI |
Maths
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | – | 86.0%#6 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | – | 70.7%#15 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 53.3%#161 | 88.9%#61 | – | – | – | 99.7%#8 | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | 74.0%#8 | – | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | – | – | 94.9%#9 | – | LiveBench |
Knowledge
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | – | 43.2%#34 | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | – | – | – | 79.3%#9 | – | LiveBench | |
| AA-Omniscience | -23.2#194 | -6.8#122 | -5.1#117 | -3.5#109 | -3.0#108 | – | 0.1#95 | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | 86.7%#39 | – | – | Vals AI | |
| LegalBench | – | – | – | – | – | 85.1%#22 | – | Vals AI | |
| CorpFin | – | – | – | – | 65.3%#29 | – | – | Vals AI | |
| TaxEval | – | – | – | – | 76.2%#6 | – | – | Vals AI |
Instruction following
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | – | – | – | 82.9%#15 | – | LiveBench | |
| IFBench | – | 59.7%#125 | 62.2%#117 | 64.4%#106 | 66.3%#96 | – | 71.2%#58 | Artificial Analysis |
Human preference
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | – | 1466#43 | – | – | LMArena | |
| EQ-Bench 4 | – | – | – | – | – | – | 1234#11 | EQ-Bench |
Multimodal
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | – | – | – | 1269#33 | – | – | LMArena | |
| MMMU-Pro | 66.7%#137 | 76.1%#66 | 76.8%#61 | 79.1%#44 | 79.5%#41 | – | 80.7%#30 | Artificial Analysis |
Long context
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| AA-LCR | 58.7%#224 | 71.3%#149 | 74.0%#120 | 77.7%#87 | 79.0%#71 | – | 83.0%#12 | Artificial Analysis |
Composite
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | – | 159.2#7 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | – | – | – | 77.9%#14 | – | LiveBench | |
| AA Intelligence Index | 22.3#148 | 27.9#96 | 32.8#69 | 34.5#56 | 38.2#46 | – | 42.3#27 | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | 59.6#12 | – | Vals AI |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| OpenAI Flex | 70 tok/s | 2.25 s | $1.00 | $6.00 | 1.1M | – |
| OpenAI Fast | 67 tok/s | 2.54 s | $4.00 | $24.00 | 1.1M | – |
| Azure (EU) | 62 tok/s | 2.75 s | $2.20 | $13.20 | 1.1M | – |
| OpenAI | 45 tok/s | 2.78 s | $2.00 | $12.00 | 1.1M | – |
| Azure | 32 tok/s | 4.57 s | $2.00 | $12.00 | 1.1M | – |
| Amazon Bedrock (US) | 31 tok/s | 1.34 s | $2.20 | $13.20 | 1.1M | – |
| Azure (US) | 30 tok/s | 7.18 s | $2.20 | $13.20 | 1.1M | – |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.220 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0044 | $0.0039 | 35.8 s |
| Summarise a 30-page report | 12,000 / 600 | $0.031 | $0.015 | 39.0 s |
| Code edit | 6,000 / 1,500 | $0.030 | $0.022 | 48.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.168 | $0.088 | 1.3 min |
| Structured extraction | 2,000 / 200 | $0.0064 | $0.0037 | 34.7 s |
See also
Data as of 9 Sept 2026. Compare these configurations.