o1 (high): benchmark scores, price and ranking
o1 (high) from OpenAI ranks #380 of 743 LLMs on the LLMs Tiger aggregate benchmark score (48.1), based on 9 benchmark results captured up to 2026-10-13.
- Organisation
- OpenAI
- Type
- Frontier model
- Open weights
- No
- Released
- 2024-12-05 (source: curated)
- Cheapest API price
- $15.00 in / $60.00 out per million tokens (via azure, observed 2026-10-13T19:47Z)
- Aggregate rank
- #380 of 743 (score 48.1)
Relative to other models, o1 (high) is strongest on WeirdML (#36 of 109) and weakest on FrontierMath T1-3 (#60 of 68).
Across the 9 benchmarks where at least ten models are scored, o1 (high) is in the top 10% on 0 and in the top half on 6.
Its cheapest listed input price, $15.00 per million tokens, is lower than that of 2% of the 427 models with a listed price.
o1 (high) benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| WeirdML | 46.1% | #36 of 109 | Epoch AI Benchmarking Hub | 2026-10-13 |
| OTIS Mock AIME | 73.3% | #72 of 189 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GPQA Diamond | 76.8% | #119 of 275 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Chess Puzzles | 15.0% | #55 of 122 | Epoch AI Benchmarking Hub | 2026-10-13 |
| DTBench | 74.7% | #68 of 146 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SimpleQA Verified | 41.1% | #34 of 69 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SimpleBench | 40.1% | #34 of 61 | Epoch AI Benchmarking Hub | 2026-10-13 |
| LMCA | 22.3% | #69 of 108 | Epoch AI Benchmarking Hub | 2026-10-13 |
| FrontierMath T1-3 | 14.7% | #60 of 68 | Epoch AI Benchmarking Hub | 2026-10-13 |
About o1 (high)
o1 (high) is a frontier model from OpenAI. It was released on 2024-12-05, according to the developer's announcement. The cheapest API listing we track charges $15.00 per million input tokens and $60.00 per million output tokens, through azure. We hold 9 benchmark results for it across 3 areas, and none is in the top tenth of its board.
Where o1 (high) stands
On our aggregate score, o1 (high) is #380 of 743 ranked models, around the middle of the table. Its score is 48.1; the scale is centred on 50, so it sits 1.9 points below the typical model. That beats 49% of the ranked field. Among the 88 ranked OpenAI models it is #49. The next one up is gpt-3.5-turbo-0613 at 48.3. The next one down is GPT-5.1 (Thinking, Medium) at 47.8.
o1 (high), benchmark by benchmark
WeirdML puts o1 (high) at 46.1%, #36 of 109. That beats 67% of the models tested, but 47.2 points short of GPT-6 Astra (max) (93.3%). The results either side are claude-sonnet-4-20250514_16K (46.1%) and claude-opus-4-1-20250805_16K (45.9%). It covers unusual machine-learning tasks. The board median is 41.2%, so o1 (high) is 4.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On OTIS Mock AIME, o1 (high) reaches 73.3%. It ranks #72 of 189, ahead of 62% of the field, and 26.7 points behind Claude Fable 5.1 (max) at 100.0%. The results either side are Gemma 4 31B IT (73.3%) and Qwen3-Max-Instruct (73.3%). It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so o1 (high) is 15.6 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
o1 (high) manages 76.8% on GPQA Diamond, good for #119 of 275. GPT-6 Astra (max) leads at 95.8%, 19.0 points further up, and 57% of models sit below. The results either side are claude-opus-4-1-20250805_27K (76.8%) and DeepSeek-R1 (May 2025) (76.3%). It covers expert-level science questions. The board median is 71.7%, so o1 (high) is 5.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Chess Puzzles puts o1 (high) at 15.0%, #55 of 122. That beats 55% of the models tested, but 57.0 points short of GPT-6 Astra (max) (72.0%). The results either side are Claude Sonnet 5 (max) (16.0%) and Claude Opus 4.6 (max) (14.0%). It covers chess tactics puzzles. The board median is 12.5%, so o1 (high) is 2.5 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On DTBench, o1 (high) reaches 74.7%. It ranks #68 of 146, ahead of 53% of the field, and 24.3 points behind Claude Opus 5.5 (max) at 98.9%. The results either side are Gemma 4 26B A4B (74.9%) and Qwen3.6 35B-A3B (73.9%). It covers dynamic tool-use benchmark. The board median is 68.8%, so o1 (high) is 5.9 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
o1 (high) manages 41.1% on SimpleQA Verified, good for #34 of 69. gemini-4-argon-high leads at 76.0%, 34.9 points further up, and 51% of models sit below. The results either side are GPT-6 Luna (max) (41.4%) and GLM-5.3 (max) (41.0%). It covers short factual questions. The board median is 41.0%, so o1 (high) is 0.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
o1 (high) lands below the middle on SimpleBench: 40.1%, #34 of 61, with 54% of the field scoring higher. Claude Fable 5 (max) leads at 81.9%; the closest result above is DeepSeek-R1 (May 2025) at 40.8%. It covers trick questions that test common sense. The board median is 41.4%, so o1 (high) is 1.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
o1 (high) lands below the middle on LMCA: 22.3%, #69 of 108, with 63% of the field scoring higher. Claude Opus 5.5 (max) leads at 68.2%; the closest result above is qwen3-30b-a3b-instruct-2507 at 22.4%. It covers language-model capability assessment. The board median is 29.2%, so o1 (high) is 7.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
FrontierMath T1-3 is a weak spot. o1 (high) scores 14.7%, #60 of 68, behind 87% of the models measured. The leader, GPT-6 Astra (max), reaches 93.7%. It covers research-level maths, tiers 1-3. The board median is 55.6%, so o1 (high) is 40.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Reasoning, maths and science: uneven, ahead of 44% of the field on average across 4 benchmarks. Its best board here is OTIS Mock AIME (#72 of 189) and its weakest is FrontierMath T1-3 (#60 of 68). General knowledge and chat: solid, ahead of 52% of the field on average across 4 benchmarks. Its best board here is WeirdML (#36 of 109) and its weakest is LMCA (#69 of 108).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so o1 (high) costs more than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Llama-3.2-3B-Instruct ($0.020, 47.2). Similar scores cost more on GPT-4 (0125) ($30.00) and GPT-4 (1106) ($30.00).
Release history
The OpenAI model released before it was GPT-4o (Nov 2024) on 2024-11-20, which scores 42.3. The next one was o1-2024-12-17 on 2024-12-17, at 51.6.
Reading these numbers
The 9 scores on this page come from 1 source: Epoch AI Benchmarking Hub (9). No official leaderboard supplied any of them, so treat them as third-party measurements. The leaderboard also lists 4 other reasoning-effort settings of this model, which redirect here. This page shows the highest.
o1 (high) is also listed at 4 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.
Models ranked nearby
- #377 Qwen3-Max-Instruct (48.2)
- #378 internlm3-8b-instruct (48.2)
- #379 kimi-k2.6-thinking (48.2)
- #381 llama3-70b-8192 (48.1)
- #382 gemini-1.5-pro-001-feb24 (48.1)
- #383 Codex + GPT-5.6 Luna (48.1)
More from OpenAI
- GPT-6 Astra (max) (#2, 64.0)
- GPT-6.1 Sol (max) (#5, 60.6)
- GPT-5.6 Sol (max) (#6, 60.1)
- GPT-6 Sol (max) (#7, 60.0)
- GPT-5.5 (xhigh) (#11, 58.0)
- GPT-5.6 Terra (max) (#12, 57.8)
Compare it on the live leaderboard → · How the aggregate score works