OpenAI, proprietary
GPT-6 Sol
GPT-6 Sol by OpenAI ranks 12th of 354 ranked models on the Noometry Index as of October 2026, with a score of 61.8. Its strongest category is math, where it ranks 7th. API pricing starts at $2 per million input tokens and $10 per million output tokens, with a 1.05M-token context window.
Last verified
Specifications
- Noometry rank
- #12 of 354
- Index score
- 61.8
- Evidence
- Confirmed 45 results
- Provider
- OpenAI
- Released
- September 22, 2026
- Weights
- Proprietary
- Reasoning
- Yes
- Context window
- 1.05M
- Max output
- 128K
- Input price
- $2 / M
- Output price
- $10 / M
- Blended price
- $4 / M
- Output speed
- Not measured
- Value
- #163 of 219
- Knowledge cutoff
- April 2026
- Input
- text, image, pdf
Category scores
Each category score combines every public result we have in that category.
- Coding 60.1
- Agentic & Tool Use 37.2
- Reasoning 74.0
- Math 87.2
- Knowledge 64.8
- Multimodal 47.6
- Multilingual 50.5
- Instruction Following 74.5
- Long Context 43.1
- Writing & Preference 71.9
| Category | Score | Rank | Results |
|---|---|---|---|
| Coding | 60.1 | #11 | 5 |
| Agentic & Tool Use | 37.2 | #36 | 2 |
| Reasoning | 74.0 | #9 | 9 |
| Math | 87.2 | #7 | 5 |
| Knowledge | 64.8 | #15 | 4 |
| Multimodal | 47.6 | #10 | 3 |
| Multilingual | 50.5 | #118 | 1 |
| Instruction Following | 74.5 | #94 | 1 |
| Long Context | 43.1 | #108 | 1 |
| Writing & Preference | 71.9 | #18 | 4 |
Strengths and weaknesses
Categories where GPT-6 Sol places highest and lowest among the models ranked in each, with its score against that category's median.
Strongest categories
Weakest categories
| Category | Score | vs median | Rank |
|---|---|---|---|
| Multilingual | 50.5 | +3.1 | #118 of 297, top 40% |
| Long Context | 43.1 | +2.1 | #108 of 296, top 37% |
| Instruction Following | 74.5 | +3.2 | #94 of 305, top 31% |
Closest competitors
The models ranked just above and below GPT-6 Sol. When scores are this close, price and speed are often the better way to choose.
| Model | Rank | Score | Blended $/M | Speed | |
|---|---|---|---|---|---|
| GPT-5.5 Pro | #8 | 64.3 | $67.50 | — | Compare |
| GPT-5.5 | #9 | 63.4 | $11.25 | 25 | Compare |
| Claude Sonnet 5.5 | #10 | 61.9 | $4 | — | Compare |
| Gemini 3.8 Flash | #11 | 61.8 | $1.50 | — | Compare |
| Claude Opus 4.8 | #13 | 60.7 | $10 | 34 | Compare |
| Gemini 3.7 Flash | #14 | 59.8 | $1.50 | — | Compare |
| Kimi K3 | #15 | 59.5 | $6 | — | Compare |
| GPT-5.4 | #16 | 59.4 | $5.63 | 12 | Compare |
Sponsored placements are available on pages like this one. Advertise on Noometry
Benchmark results
Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.
Coding
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| DeepSWE | 65.3% | high | Model card (self-reported) | 2026-09-22 | |
| DeepSWE | 37.2% | low | Model card (self-reported) | 2026-09-22 | |
| DeepSWE | 68.8% | #9 of 29, top 32% | max | Model card (self-reported) | 2026-09-22 |
| DeepSWE | 56.6% | medium | Model card (self-reported) | 2026-09-22 | |
| DeepSWE | 66.6% | xhigh | Model card (self-reported) | 2026-09-22 | |
| FrontierCode | 49.3% | #8 of 37, top 22% | max | Epoch AI | |
| FrontierCode | 47.7% | high | Model card (self-reported) | 2026-09-22 | |
| FrontierCode | 37.3% | low | Model card (self-reported) | 2026-09-22 | |
| FrontierCode | 49.3% | max | Model card (self-reported) | 2026-09-22 | |
| FrontierCode | 45.9% | medium | Model card (self-reported) | 2026-09-22 | |
| FrontierCode | 48.4% | xhigh | Model card (self-reported) | 2026-09-22 | |
| LMArena WebDev | 1688 | #7 of 113, top 7% | LMArena | 2026-10-08 | |
| SciCode | 54.9% | high | Epoch AI | ||
| SciCode | 50.2% | low | Epoch AI | ||
| SciCode | 57.6% | #15 of 121, top 13% | max | Epoch AI | |
| SciCode | 53.8% | medium | Epoch AI | ||
| SciCode | 47.3% | none | Epoch AI | ||
| SciCode | 55.1% | xhigh | Epoch AI | ||
| LMArena Coding | 1447 | #89 of 294, top 31% | LMArena | 2026-10-08 | |
| ALE-Bench | 2,462 | #2 of 105, top 2% | max | Epoch AI |
Agentic & Tool Use
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| APEX-Agents | 54.3% | #21 of 49, top 43% | max | Epoch AI | |
| GDP.pdf | 26.4% | #8 of 36, top 23% | max | Epoch AI | |
| GDP.pdf | 28% | high | Model card (self-reported) | 2026-09-29 | |
| GDP.pdf | 21.8% | low | Model card (self-reported) | 2026-09-29 | |
| GDP.pdf | 24.8% | max | Model card (self-reported) | 2026-09-29 | |
| GDP.pdf | 25.4% | medium | Model card (self-reported) | 2026-09-29 | |
| GDP.pdf | 23.8% | xhigh | Model card (self-reported) | 2026-09-29 | |
| Vending-Bench 2 | 14,428 | #2 of 60, top 4% | Epoch AI |
Reasoning
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| ARC-AGI-2 | 68.9% | high | Epoch AI | ||
| ARC-AGI-2 | 31.5% | low | Epoch AI | ||
| ARC-AGI-2 | 89.6% | #7 of 83, top 9% | max | Epoch AI | |
| ARC-AGI-2 | 57.8% | medium | Epoch AI | ||
| ARC-AGI-2 | 1.7% | none | Epoch AI | ||
| ARC-AGI-2 | 78.1% | xhigh | Epoch AI | ||
| NYT Connections (extended) | 90.1% | #17 of 91, top 19% | high reasoning | Lech Mazur benchmarks | |
| ARC-AGI-1 | 91% | high | Epoch AI | ||
| ARC-AGI-1 | 72.2% | low | Epoch AI | ||
| ARC-AGI-1 | 95.5% | #14 of 83, top 17% | max | Epoch AI | |
| ARC-AGI-1 | 83.7% | medium | Epoch AI | ||
| ARC-AGI-1 | 29.3% | none | Epoch AI | ||
| ARC-AGI-1 | 92.7% | xhigh | Epoch AI | ||
| CritPt | 25.4% | high | Epoch AI | ||
| CritPt | 16.3% | low | Epoch AI | ||
| CritPt | 30.9% | #7 of 134, top 6% | max | Epoch AI | |
| CritPt | 24.6% | medium | Epoch AI | ||
| CritPt | 4% | none | Epoch AI | ||
| CritPt | 28% | xhigh | Epoch AI | ||
| EBR-Bench | 53.3% | #5 of 24, top 21% | max | Epoch AI | 2026-09-22 |
| LMArena Hard Prompts | 1418 | #107 of 297, top 37% | LMArena | 2026-10-08 | |
| Mystery Game Puzzles | 56% | #9 of 74, top 13% | max | Epoch AI | 2026-09-22 |
| DTBench | 97.3% | #6 of 151, top 4% | max | Epoch AI | |
| LMCA | 59.1% | #7 of 125, top 6% | max | Epoch AI | |
| Epoch Capabilities Index | 162.72 | #7 of 213, top 4% | Epoch AI | 2026-09-22 |
Math
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| FrontierMath (Tiers 1-3) | 89.8% | #5 of 81, top 7% | max | Epoch AI | 2026-09-22 |
| FrontierMath Tier 4 | 90% | #5 of 63, top 8% | max | Epoch AI | 2026-09-22 |
| OTIS Mock AIME 2024-2025 | 100% | #10 of 173, top 6% | max | Epoch AI | 2026-09-22 |
| ProofBench | 83% | #11 of 77, top 15% | Epoch AI | ||
| LMArena Math | 1402 | #123 of 285, top 44% | LMArena | 2026-10-08 |
Knowledge
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| GPQA Diamond | 94.3% | #8 of 186, top 5% | max | Epoch AI | 2026-09-22 |
| SimpleQA Verified | 60.7% | #14 of 77, top 19% | max | Epoch AI | 2026-09-22 |
| Vectara Hallucination Rate (lower is better) | 6.5% | #25 of 96, top 27% | Vectara Hallucination Leaderboard | ||
| LMArena Expert | 1439 | #86 of 273, top 32% | LMArena | 2026-10-08 |
Multimodal
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Vision | 1245 | #60 of 122, top 50% | LMArena | 2026-10-09 | |
| Blueprint-Bench 2 | 36.9% | #7 of 31, top 23% | Epoch AI | ||
| Furniture Assembly | 58.3% | #7 of 31, top 23% | max | Epoch AI | 2026-09-23 |
Multilingual
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Non-English | 1385 | #118 of 297, top 40% | LMArena | 2026-10-08 | |
| LMArena Chinese | 1405 | #135 of 285, top 48% | LMArena | 2026-10-08 | |
| LMArena French | 1410 | #111 of 223, top 50% | LMArena | 2026-10-08 | |
| LMArena German | 1390 | #103 of 231, top 45% | LMArena | 2026-10-08 | |
| LMArena Japanese | 1385 | #72 of 211, top 35% | LMArena | 2026-10-08 | |
| LMArena Korean | 1341 | #109 of 213, top 52% | LMArena | 2026-10-08 | |
| LMArena Russian | 1401 | #103 of 283, top 37% | LMArena | 2026-10-08 | |
| LMArena Spanish | 1384 | #124 of 226, top 55% | LMArena | 2026-10-08 |
Instruction Following
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Instruction Following | 1412 | #84 of 298, top 29% | LMArena | 2026-10-08 |
Long Context
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Longer Query | 1411 | #105 of 291, top 37% | LMArena | 2026-10-08 |
Writing & Preference
| Benchmark | Score | Position | Setting | Source | Date |
|---|---|---|---|---|---|
| LMArena Text | 1395 | #124 of 297, top 42% | LMArena | 2026-10-08 | |
| LMArena Creative Writing | 1378 | #105 of 295, top 36% | LMArena | 2026-10-08 | |
| EQ-Bench Creative Writing | 2125 | #4 of 115, top 4% | EQ-Bench | ||
| LMArena Multi-Turn | 1412 | #105 of 295, top 36% | LMArena | 2026-10-08 |
API pricing by provider
| Route | Input $/M | Output $/M | Cached input $/M | Checked |
|---|---|---|---|---|
| azure | $2 | $10 | $0.20 | 2026-10-10 |
| bedrock | $2 | $10 | $0.20 | 2026-10-10 |
| openai | $2 | $10 | $0.20 | 2026-10-10 |
| openrouter | $2 | $10 | $0.20 | 2026-10-10 |
Compare GPT-6 Sol
- GPT-6 Sol vs GPT-5.6 Sol
- GPT-6 Sol vs Gemini 3.8 Flash
- GPT-6 Sol vs Claude Opus 4.8
- GPT-6 Sol vs Claude Sonnet 5.5
- GPT-6 Sol vs Gemini 3.7 Flash
- GPT-6 Sol vs GPT-5.5
- GPT-6 Sol vs Kimi K3
- GPT-6 Sol vs Claude Fable 5.1
- GPT-6 Sol vs Grok 4.6
- GPT-6 Sol vs Qwen3.8 Max
- GPT-6 Sol vs GLM-5.3
- GPT-6 Sol vs Muse Spark 1.3
- GPT-6 Sol vs DeepSeek V4 Pro
- GPT-6 Sol vs MiMo-V2.6-Pro
Other OpenAI models
- GPT-6 Astra70.8
- GPT-6.1 Sol65.6
- GPT-5.6 Sol65.0
- GPT-5.5 Pro64.3
- GPT-5.563.4
- GPT-5.459.4
- GPT-5.6 Terra59.2
- GPT-5.4 Pro58.9
Frequently asked questions
How good is GPT-6 Sol?
GPT-6 Sol by OpenAI ranks 12th of 354 ranked models on the Noometry Index as of October 2026, with a score of 61.8. Its strongest category is math, where it ranks 7th. API pricing starts at $2 per million input tokens and $10 per million output tokens, with a 1.05M-token context window.
How much does GPT-6 Sol cost?
GPT-6 Sol costs $2 per million input tokens and $10 per million output tokens on OpenAI's own API, with cached input at $0.20.
What is GPT-6 Sol's context window?
GPT-6 Sol accepts up to 1.05M tokens of input and can write up to 128K tokens in one response.
Is GPT-6 Sol open source?
No. GPT-6 Sol is proprietary and available only through OpenAI's API and partner platforms.
What are GPT-6 Sol's strengths and weaknesses?
Relative to other ranked models, GPT-6 Sol places best in math, reasoning, coding and lowest in multilingual, long context, instruction following.
What is GPT-6 Sol best at?
Its best category is math, where it ranks 7th on Noometry.