OpenAIGPT-5.6 Sol
Best configuration ranks #12 of 610 on the BenchLeader Index at 68.0 ±6.3 (high reasoning effort). Last measured 9 Jul 2026. Released 9 Jul 2026.
- Blended price
- $8.00/M
- $4.00 in · $20.00 out
- Output speed
- 59 tok/s
- First answer
- 31 s
- first token 5.46 s
- Context
- 1M
- Overall index68
- Reasoning63
- Coding72
- Agents & tools88
- Knowledge74
- Instruction following68
- Multimodal67
- Long context67
- Composite83
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 | 56.1 | #164 | 65 tok/s | 1.28 s | $0.0076 | Coding 54 · Composite 65 · Knowledge 65 · Long context 57 · Maths 54 · Multimodal 56 · Reasoning 60 |
| low | 64.6 | #32 | 58 tok/s | 2.90 s | $0.0076 | Agents & tools 86 · Coding 65 · Composite 72 · Instruction following 65 · Knowledge 73 · Long context 66 · Maths 66 · Multimodal 66 · Reasoning 56 |
| medium | 66.0 | #21 | 64 tok/s | 7.87 s | $0.0076 | Agents & tools 88 · Coding 66 · Composite 79 · Instruction following 68 · Knowledge 73 · Long context 67 · Multimodal 66 · Reasoning 59 |
| highbest | 68.0 | #12 | 59 tok/s | 31 s | $0.0076 | Agents & tools 88 · Coding 72 · Composite 83 · Instruction following 68 · Knowledge 74 · Long context 67 · Multimodal 67 · Reasoning 63 |
| xhigh | 67.5 | #14 | 63 tok/s | 54 s | $0.0076 | Agents & tools 72 · Coding 68 · Composite 85 · Human preference 68 · Instruction following 69 · Knowledge 74 · Long context 68 · Multimodal 65 · Reasoning 65 |
| max | 65.5 | #26 | 68 tok/s | 140 s | $0.0084 | Agents & tools 59 · Coding 70 · Composite 70 · Knowledge 65 · Maths 74 · Reasoning 66 |
| default | 65.3 | #28 | 68 tok/s | 140 s | $0.0076 | Agents & tools 68 · Coding 64 · Composite 88 · Human preference 58 · Instruction following 71 · Knowledge 74 · Long context 69 · Multimodal 68 · Reasoning 62 |
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 | 82.8%#98 | 89.9%#39 | – | – | – | 93.5%#12 | – | Epoch AI Benchmarking Hub | |
| SimpleBench | – | – | – | – | 64.8%#17 | – | 64.8%#17 | SimpleBench | |
| LMArena Hard Prompts | – | – | – | – | 1507#17 | – | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | – | – | – | 91.7%#3 | – | LiveBench | |
| GPQA Diamond (AA)not in index | 79.0%#197 | 89.8%#68 | 92.6%#32 | 92.8%#28 | 93.1%#24 | – | 94.1%#7 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 16.7%#181 | 39.4%#59 | 42.2%#43 | 46.0%#26 | 47.3%#22 | – | 49.5%#14 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | – | 95.2%#2 | – | Vals AI | |
| EnigmaEval | – | – | – | 37.1%#2 | – | – | – | Scale AI SEAL | |
| Kagi LLM Benchmark | – | – | – | – | – | – | 67.0%#35 | Kagi LLM Benchmark | |
| ARC-AGI-1 | – | 74.5%#94 | 92.5%#35 | 97.0%#12 | 97.5%#6 | 96.5%#13 | – | ARC Prize | |
| ARC-AGI-2 | – | 42.5%#78 | 67.1%#43 | 85.4%#16 | 90.0%#7 | 92.5%#3 | – | ARC Prize | |
| ARC-AGI-3 | – | 0.3%#26 | 1.1%#19 | 2.2%#16 | 7.0%#15 | 7.8%#14 | – | ARC Prize |
Coding
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SciCode | 47.1%#76 | 55.4%#24 | 56.5%#13 | 56.9%#10 | 56.0%#19 | 56.1%#17 | – | SciCode | |
| WeirdML | – | – | – | 88.8%#8 | – | 87.0%#10 | – | WeirdML | |
| FrontierCode | – | – | – | – | – | – | 47.5%#6 | Cognition | |
| LMArena Coding | – | – | – | – | 1529#14 | – | – | LMArena | |
| LMArena WebDev | – | – | – | – | 1617#12 | – | – | LMArena | |
| LiveBench Codingnot in index | – | – | – | – | – | 83.9%#3 | – | LiveBench | |
| SciCode (AA)not in index | – | 56.4%#22 | 57.4%#13 | 57.8%#12 | 57.1%#16 | – | 57.1%#16 | Artificial Analysis | |
| LiveCodeBench | – | – | – | – | – | 82.6%#50 | – | Vals AI | |
| IOI | – | – | – | – | – | 86.7%#2 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | – | 96.2%#3 | – | Vals AI |
Agents & tools
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | – | – | – | 37.3%#38 | – | Terminal-Bench | |
| OSWorld-Verified 2.0 | – | – | – | – | – | 27.3%#4 | – | OSWorld | |
| APEX-Agents | – | – | – | – | 37.7%#14 | 39.9%#10 | 37.7%#14 | Mercor | |
| LMArena Agent | – | – | – | – | 7.7#6 | – | – | LMArena | |
| LiveBench Agentic Codingnot in index | – | – | – | – | – | 56.2%#21 | – | LiveBench | |
| Terminal-Bench Hard | – | 60.6%#7 | 62.9%#2 | 62.1%#5 | 61.4%#6 | – | 65.9%#1 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | – | 76.0%#130 | 81.0%#115 | 83.3%#106 | 84.8%#95 | – | 85.1%#92 | Artificial Analysis | |
| Terminal-Bench 2.1 (Vals) | – | – | – | – | – | 85.8%#2 | – | Vals AI | |
| HiL-Bench | – | – | – | – | – | – | 32.3%#11 | Scale AI SEAL |
Maths
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | – | 89.1%#3 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | – | 82.9%#8 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 68.9%#126 | 95.6%#29 | – | – | – | 100.0%#1 | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | – | 83.0%#6 | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | – | – | 96.2%#3 | – | LiveBench |
Knowledge
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | – | 69.7%#5 | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | – | – | – | 79.8%#5 | – | LiveBench | |
| AA-Omniscience | 1.1#85 | 18.9#45 | 19.4#44 | 20.4#42 | 21.0#39 | – | 22.0#36 | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | – | 89.1%#16 | – | Vals AI | |
| LegalBench | – | – | – | – | – | 87.0%#7 | – | Vals AI | |
| CorpFin | – | – | – | – | – | 64.4%#38 | – | Vals AI | |
| TaxEval | – | – | – | – | – | 74.8%#28 | – | Vals AI | |
| PRBench Finance | – | – | – | – | – | 50.5%#8 | – | Scale AI SEAL | |
| PRBench Legal | – | – | – | – | – | 50.5%#6 | – | Scale AI SEAL |
Instruction following
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | – | – | – | 87.7%#6 | – | LiveBench | |
| IFBench | – | 66.5%#93 | 69.6%#72 | 69.2%#73 | 71.0%#62 | – | 72.7%#46 | Artificial Analysis |
Human preference
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | – | 1483#17 | – | – | LMArena | |
| EQ-Bench 4 | – | – | – | – | – | – | 1250#9 | EQ-Bench |
Multimodal
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | – | – | – | 1277#30 | – | – | LMArena | |
| MMMU-Pro | 71.9%#104 | 81.0%#29 | 81.4%#26 | 81.8%#24 | 82.7%#18 | – | 83.4%#16 | Artificial Analysis |
Long context
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| AA-LCR | 62.3%#209 | 78.0%#82 | 80.3%#45 | 81.7%#30 | 82.3%#21 | – | 84.0%#6 | Artificial Analysis |
Composite
| Benchmark | no reasoning | low | medium | high | xhigh | max | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | – | 161.8#6 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | – | – | – | 81.0%#5 | – | LiveBench | |
| AA Intelligence Index | 28.3#91 | 33.8#63 | 39.5#39 | 42.5#26 | 44.1#23 | – | 47.1#14 | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | 63.7#6 | – | 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 Fast | 61 tok/s | 2.10 s | $4.00 | $20.00 | 1.1M | – |
| Azure (EU) | 59 tok/s | 5.15 s | $5.50 | $33.00 | 1.1M | – |
| OpenAI | 41 tok/s | 4.14 s | $2.00 | $10.00 | 1.1M | – |
| Amazon Bedrock (US) | 38 tok/s | 10 s | $4.40 | $22.00 | 1.1M | – |
| OpenAI Flex | 37 tok/s | 9.34 s | $1.00 | $5.00 | 1.1M | – |
| Azure | 29 tok/s | 5.46 s | $5.00 | $30.00 | 1.1M | – |
| Azure (US) | 18 tok/s | 16 s | $5.50 | $33.00 | 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 $0.440 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0076 | $0.0065 | 36.5 s |
| Summarise a 30-page report | 12,000 / 600 | $0.060 | $0.028 | 41.6 s |
| Code edit | 6,000 / 1,500 | $0.054 | $0.038 | 56.8 s |
| Agentic coding session | 60,000 / 4,000 | $0.320 | $0.160 | 1.7 min |
| Structured extraction | 2,000 / 200 | $0.012 | $0.0067 | 34.8 s |
See also
Data as of 9 Sept 2026. Compare these configurations.