o1-preview
Better value: Nemotron 3 Nano 30B A3B scores 47.4 for $0.088/M, 328× cheaper.
173 cheaper ranked models score higher: Claude Fable 5.2 (70.2, $16.00), Gemini 4 Carbon (68.6, $4.00), GPT-6.2 Astra (68.3, $16.00) and 170 more.
As of 22 Nov 2026, o1-preview by OpenAI scores 48.4 on the BenchLeader Index (650 ranked configurations), with its best results in human preference (56) and weakest in agents & tools (38). API pricing: $16.50 per 1M input tokens and $66.00 per 1M output tokens ($28.88 blended at 3:1), list price on 22 Nov 2026. Released 12 Sep 2024.
Key facts (as of 22 Nov 2026)
- o1-preview API pricing: $16.50 per 1M input tokens and $66.00 per 1M output tokens ($28.88 blended at 3:1) (list price, 22 Nov 2026).
- o1-preview was released on 12 Sep 2024 (OpenAI).
In detail. o1-preview is an OpenAI proprietary reasoning model, released 12 Sep 2024. Its best configuration ranks #406 of 650 on the BenchLeader Index at 48.4 ±5.4, in the lower half. The ± is the point: 301 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 human preference (56) and lowest in agents & tools (38). At $28.88 per million tokens blended it is among the most expensive ranked models. It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 19 Nov 2026.
- Blended price
- $28.88/M
- $16.50 in · $66.00 out
- via Artificial Analysis
- Output speed
- –
- First answer
- –
- Context
- 128k
- via Artificial Analysis
- Cost per task
- –
- Tokens per task
- –
- Released
- 12 Sep 2024
- Overall index48
- Reasoning43
- Coding51
- Agents & tools38
- Maths53
- Human preference56
Embed this rank
More badge styles[](https://www.benchleader.com/models/o1-0912)Paste it at the top of README.md, beside your other badges. The badge redraws with every data update and links back to this page.
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 |
|---|---|---|---|---|---|---|
| high | – | – | – | – | $0.026 | Knowledge 59 |
| not statedbest | 48.4 | #406 | – | – | $0.026 | Agents & tools 38 · Coding 51 · Human preference 56 · Maths 53 · Reasoning 43 |
Benchmark results
See all LLM benchmarks. 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 | not stated | Source | Trend |
|---|---|---|---|
| GPQA Diamond | 50.3%#237 | Epoch AI Benchmarking Hub | |
| SimpleBench | 41.7%#81 | SimpleBench | |
| Arena Hard Prompts | 1396#195 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 76.5%#242 | Artificial Analysis | |
| Arena Expertnot in index | 1383#200 | Arena (LMArena) |
Coding
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| WeirdML | 47.6%#77 | WeirdML | |
| Arena Coding | 1415#202 | Arena (LMArena) | |
| SWE-bench Verified (any scaffold)not in index | 64.6%#23 | SWE-bench |
Agents & tools
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Cybench | 10.0%#15 | Cybench | |
| METR Time Horizonsnot in index | 45.1%#30 | METR | |
| METR 50% Time Horizonnot in index | 20 min#17 | METR |
Maths
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| OTIS Mock AIME | 31.1%#223 | Epoch AI Benchmarking Hub | |
| MATH Level 5 | 81.7%#34 | Epoch AI Benchmarking Hub | |
| Arena Maths | 1386#181 | Arena (LMArena) | |
| AIME 2025 (AA)not in index | 79.7%#68 | Artificial Analysis |
Instruction following
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Instruction Followingnot in index | 1381#179 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1389#182 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1366#168 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1390#185 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1382#193 | Arena (LMArena) |
Composite
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Epoch Capabilities Indexnot in index | 133.6#146 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index v4.3.2not in index | 11.4#370 | Artificial Analysis | |
| AA Progress Index v1not in index | 226.1#371 | Artificial Analysis |
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 $8.25 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.026 | $0.024 | – |
| Summarize a 30-page report | 12,000 / 600 | $0.238 | $0.163 | – |
| Code edit | 6,000 / 1,500 | $0.198 | $0.161 | – |
| Agentic coding session | 60,000 / 4,000 | $1.25 | $0.883 | – |
| Structured extraction | 2,000 / 200 | $0.046 | $0.034 | – |
More from OpenAI
See also
Embed or cite
Show where o1-preview ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page.
[](https://www.benchleader.com/models/o1-0912)Cite this
The ranking changes daily, so a reference carries the data date. APA style:
BenchLeader. (2026, November 22). o1-preview on the BenchLeader Index (method v1.5) [Data set]. Retrieved November 22, 2026, from https://www.benchleader.com/models/o1-0912
BibTeX:
@misc{benchleader_o1_0912_20261122,
author = {{BenchLeader}},
title = {{o1-preview on the BenchLeader Index (method v1.5)}},
year = {2026},
month = nov,
howpublished = {\url{https://www.benchleader.com/models/o1-0912}},
note = {Data as of 2026-11-22}
}Every badge style, the embeddable top-10 widget and how to cite the Index
Questions people ask
- What are o1-preview's benchmark scores?
- As of 22 Nov 2026, o1-preview scores 48.4 ±5.4 on the BenchLeader Index, strongest in human preference (56) and weakest in agents & tools (38). That rests on independent benchmark results only, and every score on this page links to its publisher.
- How does o1-preview rank?
- o1-preview is #406 of 650 ranked configurations on 22 Nov 2026. 301 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Fable 5.2 (max effort) at 70.2.
- How much does o1-preview cost per 1M tokens?
- As of 22 Nov 2026, o1-preview costs $16.50 per 1M input tokens and $66.00 per 1M output tokens, $28.88 blended (three input tokens per output token).
- When was o1-preview released?
- OpenAI released o1-preview on 12 Sep 2024. BenchLeader has results for it at 2 reasoning-effort settings.
Data as of 22 Nov 2026. Compare these configurations.
Cite as: BenchLeader, “o1-preview: benchmarks, pricing, speed and rank”, data as of 22 Nov 2026, https://www.benchleader.com/models/o1-0912.