OpenAIReasoning modelGPT-6 Sol
GPT-6 Sol is an OpenAI proprietary reasoning model, released 22 Sept 2026. Its best configuration (max reasoning effort) ranks #31 of 751 on the BenchLeader Index at 66.3 ±3.2, in the upper half. The ± is the point: 88 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in composite (73) and lowest in agents & tools (64). At $4.00 per million tokens blended it is among the most expensive ranked models. Output speed of 93 tokens per second puts it faster than most, with a first answer in 140.8 s. It has been measured at 7 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 2 Oct 2026.
- Blended price
- $4.00/M
- $2.00 in · $10.00 out
- Output speed
- 93 tok/s
- measured by Artificial Analysis
- First answer
- 141 s
- first token 3.41 s
- Context
- 1.1M
- Full answer
- 146 s
- median, reasoning included
- Cost per run
- $1.04
- one full Intelligence Index run
- Released
- 22 Sept 2026
- Overall index66
- Reasoning72
- Coding68
- Agents & tools64
- Maths70
- Knowledge70
- Human preference65
- Multimodal66
- Long context67
- Composite73
Versions
OpenAI has shipped 3 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| NewGPT-6.1 Solmax | 29 Sept 2026 | 69.5 | #10 |
| GPT-6 Solmaxthis page | 22 Sept 2026 | 66.3 | #31 |
| GPT-5.6 Solmax | 9 Jul 2026 | 67.5 | #24 |
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. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 53.2 | #275 | 79 tok/s | 1.01 s | $0.0038 | Coding 52 · Composite 62 · Knowledge 61 · Long context 57 · Multimodal 53 · Reasoning 40 |
| low | 60.1 | #111 | 78 tok/s | 1.91 s | $0.0038 | Coding 56 · Composite 69 · Knowledge 74 · Long context 65 · Multimodal 63 · Reasoning 55 |
| medium | 63.0 | #71 | 53 tok/s | 3.41 s | $0.0038 | Coding 61 · Composite 75 · Knowledge 74 · Long context 67 · Multimodal 64 · Reasoning 64 |
| high | 64.2 | #56 | 87 tok/s | 18 s | $0.0038 | Coding 62 · Composite 78 · Knowledge 74 · Long context 67 · Multimodal 65 · Reasoning 67 |
| xhigh | 64.8 | #47 | 90 tok/s | 51 s | $0.0038 | Coding 62 · Composite 80 · Knowledge 74 · Long context 66 · Multimodal 66 · Reasoning 70 |
| maxbest | 66.3 | #31 | 93 tok/s | 141 s | $0.0038 | Agents & tools 64 · Coding 68 · Composite 73 · Human preference 65 · Knowledge 70 · Long context 67 · Maths 70 · Multimodal 66 · Reasoning 72 |
| not stated | – | – | 53 tok/s | 3.41 s | $0.0038 | Knowledge 38 · Maths 72 |
What more thinking costs
Turning the effort up buys index points and multiplies the bill. Here max costs 7.9× low for +6.2 on the index — 2.1 points per doubling of spend. Compare that with other models.
How its index has moved
23 Sept 2026 to 3 Oct 2026- max
- xhigh
- high
- medium
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
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 | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | – | – | – | – | – | 94.3%#8 | – | Epoch AI Benchmarking Hub | |
| LMArena Hard Prompts | – | – | – | – | – | 1484#57 | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | – | – | – | 88.7%#21 | – | LiveBench | |
| Humanity's Last Exam (AA)not in index | 18.4%#212 | 34.9%#110 | 41.0%#73 | 44.1%#48 | 46.3%#41 | 47.9%#32 | – | Artificial Analysis | |
| ARC-AGI-1 | 29.3%#176 | 72.2%#118 | 83.7%#98 | 91.0%#60 | 92.7%#46 | 95.5%#29 | – | ARC Prize | |
| ARC-AGI-2 | 1.7%#178 | 31.5%#111 | 57.8%#82 | 68.9%#55 | 78.1%#39 | 89.6%#16 | – | ARC Prize | |
| ARC-AGI-3 | 1.2%#26 | 1.8%#24 | 4.9%#21 | 5.9%#20 | 9.5%#17 | 23.0%#15 | – | ARC Prize | |
| CritPt | 4.0%#149 | 16.3%#78 | 24.6%#44 | 25.4%#40 | 28.0%#26 | 30.9%#11 | – | Artificial Analysis | |
| MysteryMechanismnot in index | – | – | – | – | – | 30.2%#12 | – | Vals AI | |
| Mystery Game Puzzlesnot in index | – | – | – | – | – | 56.0%#8 | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | – | – | – | – | – | 59.1%#23 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | – | – | – | – | – | 97.3%#8 | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SciCode | 47.3%#102 | 50.2%#84 | 53.8%#54 | 54.9%#43 | 55.1%#40 | 57.6%#16 | – | SciCode | |
| FrontierCode | – | – | – | – | – | 49.3%#9 | – | Cognition | |
| LMArena Coding | – | – | – | – | – | 1521#36 | – | LMArena | |
| LMArena WebDev | – | – | – | – | – | 1689#8 | – | LMArena | |
| LiveBench Codingnot in index | – | – | – | – | – | 81.8%#13 | – | LiveBench | |
| SciCode (AA)not in index | 47.3%#115 | 50.2%#98 | 53.8%#66 | 54.9%#53 | 55.1%#48 | 57.6%#23 | – | Artificial Analysis | |
| IOI | – | – | – | – | – | 82.6%#10 | – | Vals AI | |
| Code Migrationnot in index | – | – | – | – | – | 57.2%#7 | – | Vals AI | |
| ProgramBenchnot in index | – | – | – | – | – | 2.0%#6 | – | Vals AI | |
| Vibe Code Bench v1.1not in index | – | – | – | – | – | 87.8%#9 | – | Vals AI | |
| ALE-Benchnot in index | – | – | – | – | – | 2462#2 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | 13.1%#65 | 9.1%#78 | 18.7%#56 | 26.3%#43 | 30.3%#40 | 43.9%#23 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| APEX-Agents | – | – | – | – | – | 54.3%#18 | – | Mercor | |
| LMArena Agent | – | – | – | – | – | 9.7#6 | – | LMArena | |
| LiveBench Agentic Codingnot in index | – | – | – | – | – | 52.9%#34 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | – | – | – | – | – | 83.2%#6 | – | Vals AI | |
| GDPval (AA)not in index | 37.7%#96 | 35.3%#106 | 42.4%#75 | 44.8%#65 | 47.8%#50 | 50.3%#42 | – | Artificial Analysis | |
| ITBench SRE (AA)not in index | – | – | – | – | – | 49.4%#7 | – | Artificial Analysis | |
| CyberBenchnot in index | – | – | – | – | – | 78.0%#1 | – | Vals AI | |
| Finance Agent v2not in index | – | – | – | – | – | 49.0%#40 | – | Vals AI | |
| Harvey's Legal Agent Benchmarknot in index | – | – | – | – | – | 1.7%#41 | – | Vals AI | |
| Legal Research Benchnot in index | – | – | – | – | – | 28.9%#41 | – | Vals AI | |
| Tax Agent Benchnot in index | – | – | – | – | – | 53.0%#38 | – | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in index | – | – | – | – | – | 44.4%#8 | – | Vals AI | |
| Terminal-Bench Sciencenot in index | – | – | – | – | – | 30.0%#6 | – | Vals AI | |
| Vending-Bench 2not in index | – | – | – | – | – | – | 14427.9#2 | Andon Labs | |
| AutomationBenchnot in index | – | – | – | – | – | 61.6%#19 | – | Artificial Analysis | |
| GDP.pdfnot in index | – | – | – | – | – | 25.2%#16 | – | Artificial Analysis | |
| AA-Briefcasenot in index | – | – | – | – | – | 1480#28 | – | Artificial Analysis | |
| FACTS Searchnot in index | – | – | – | – | – | – | 86.2%#4 |
Maths
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | – | 89.8%#5 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | – | 90.0%#8 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | – | – | – | – | – | 100.0%#1 | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | – | – | 83.0%#9 | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | – | – | 96.4%#8 | – | LiveBench | |
| LMArena Maths | – | – | – | – | – | 1451#70 | – | LMArena |
Knowledge
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | – | 60.7%#14 | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | – | – | – | 81.2%#5 | – | LiveBench | |
| AA-Omniscience | -0.8#129 | 26.5#46 | 27.0#43 | 26.8#44 | 26.7#45 | 27.1#42 | – | Artificial Analysis | |
| PRBench Legal | – | – | – | – | – | – | 38.6%#30 | Scale AI SEAL | |
| BioMysteryBenchnot in index | – | – | – | – | – | 74.8%#7 | – | Vals AI | |
| Excel Modeling Benchmarknot in index | – | – | – | – | – | 71.5%#10 | – | Vals AI | |
| MedCodenot in index | – | – | – | – | – | 47.1%#34 | – | Vals AI | |
| MedScribenot in index | – | – | – | – | – | 82.0%#48 | – | Vals AI | |
| Public Benefits Benchnot in index | – | – | – | – | – | 56.6%#32 | – | Vals AI | |
| EBR-benchnot in index | – | – | – | – | – | 53.3%#5 | – | Epoch AI Benchmarking Hub | |
| GDP.pdfnot in index | – | – | – | – | – | 26.4%#10 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 45.2%#89 | 51.3%#57 | 53.5%#46 | 53.7%#44 | 53.9%#43 | 54.5%#40 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 16.0%#313 | 49.3%#105 | 43.2%#134 | 41.9%#138 | 41.1%#141 | 39.9%#146 | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | – | – | – | 85.3%#15 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | – | – | – | 68.6%#33 | – | LiveBench | |
| LMArena Instruction Followingnot in index | – | – | – | – | – | 1453#60 | – | LMArena |
Human preference
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | – | – | – | 1457#64 | – | LMArena | |
| LMArena Creative Writingnot in index | – | – | – | – | – | 1442#57 | – | LMArena | |
| LMArena Multi-turnnot in index | – | – | – | – | – | 1469#63 | – | LMArena | |
| LMArena Longer Queriesnot in index | – | – | – | – | – | 1467#63 | – | LMArena |
Multimodal
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| MMMU-Pro | 70.3%#138 | 80.2%#48 | 80.4%#46 | 82.0%#35 | 82.5%#31 | 83.0%#28 | – | Artificial Analysis | |
| SAGEnot in index | – | – | – | – | – | 44.8%#43 | – | Vals AI | |
| Blueprint-Bench 2not in index | – | – | – | – | – | – | 36.9%#5 | Epoch AI Benchmarking Hub |
Long context
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| AA-LCR | 64.0%#241 | 79.3%#89 | 82.3%#35 | 83.7%#15 | 81.3%#50 | 83.7%#15 | – | Artificial Analysis | |
| MLCRnot in index | – | – | – | – | – | 16.1%#25 | – | Artificial Analysis |
Composite
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| LiveBench | – | – | – | – | – | 79.3%#13 | – | LiveBench | |
| AA Intelligence Index | 28.5#117 | 34.1#85 | 39.8#57 | 42.4#44 | 44.2#38 | 47.6#25 | – | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | 57.5#11 | – | Vals AI |
Safety & honesty
| Benchmark | no reasoning | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|---|
| FORTRESSnot in index | – | – | – | – | – | – | 13.5%#57 | Scale AI SEAL |
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 |
|---|---|---|---|---|---|---|
| Azure | 128 tok/s | 0.92 s | $2.20 | $11.00 | 1.1M | – |
| OpenAI | 90 tok/s | 1.42 s | $4.00 | $20.00 | 1.1M | – |
| OpenAI | 77 tok/s | 0.89 s | $1.00 | $5.00 | 1.1M | – |
| OpenAI | 53 tok/s | 1.28 s | $2.00 | $10.00 | 1.1M | – |
| Azure | 44 tok/s | 2.28 s | $2.00 | $10.00 | 1.1M | – |
| Amazon Bedrock | 33 tok/s | 8.01 s | $2.20 | $11.00 | 1.1M | – |
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 $0.200 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0038 | $0.0033 | 2.4 min |
| Summarise a 30-page report | 12,000 / 600 | $0.030 | $0.014 | 2.5 min |
| Code edit | 6,000 / 1,500 | $0.027 | $0.019 | 2.6 min |
| Agentic coding session | 60,000 / 4,000 | $0.160 | $0.079 | 3.1 min |
| Structured extraction | 2,000 / 200 | $0.0060 | $0.0033 | 2.4 min |
See also
Data as of 4 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “GPT-6 Sol: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/gpt-6-sol, data as of 4 Oct 2026.