OpenAIReasoning modelo3-mini
Best configuration ranks #307 of 610 on the BenchLeader Index at 49.5 ±5.2. Last measured 2 Sept 2026. Released 31 Jan 2025.
- Blended price
- $1.93/M
- $1.10 in · $4.40 out
- Output speed
- 213 tok/s
- First answer
- 5.59 s
- first token 5.49 s
- Context
- 200k
- Overall index50
- Reasoning53
- Coding58
- Agents & tools40
- Human preference52
- Composite45
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 |
|---|---|---|---|---|---|---|
| low | 40.4 | #507 | 213 tok/s | 5.59 s | $0.0018 | Coding 34 · Maths 34 · Reasoning 40 |
| medium | 45.6 | #386 | 213 tok/s | 5.59 s | $0.0018 | Agents & tools 46 · Long context 41 · Maths 46 · Reasoning 42 |
| high | 47.8 | #344 | 213 tok/s | 21 s | $0.0018 | Agents & tools 39 · Coding 47 · Composite 43 · Human preference 54 · Instruction following 66 · Knowledge 42 · Long context 47 · Maths 50 · Reasoning 42 |
| defaultbest | 49.5 | #307 | 213 tok/s | 5.59 s | $0.0018 | Agents & tools 40 · Coding 58 · Composite 45 · Human preference 52 · Reasoning 53 |
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 | low | medium | high | default | Source |
|---|---|---|---|---|---|
| GPQA Diamond | 68.2%#160 | 74.3%#140 | 77.0%#127 | – | Epoch AI Benchmarking Hub |
| SimpleBench | – | – | 22.8%#75 | – | SimpleBench |
| LMArena Hard Prompts | – | – | 1401#152 | 1371#174 | LMArena |
| GPQA Diamond (AA)not in index | – | – | 77.3%#214 | 74.8%#245 | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | – | – | 12.0%#220 | 7.9%#285 | Artificial Analysis |
| GPQA Diamond (Vals)not in index | – | – | 75.5%#77 | – | Vals AI |
| ARC-AGI-1 | 14.5%#162 | 22.3%#153 | 34.5%#138 | – | ARC Prize |
| ARC-AGI-2 | 0.0%#174 | 2.1%#141 | 3.0%#135 | – | ARC Prize |
Coding
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| SciCode | – | – | 39.8%#112 | – | SciCode |
| WeirdML | – | – | 43.7%#86 | – | WeirdML |
| GSO-Bench | 1.3%#29 | – | 1.3%#29 | – | GSO-Bench |
| LMArena Coding | – | – | 1435#146 | 1417#159 | LMArena |
| SciCode (AA)not in index | – | – | 42.8%#103 | – | Artificial Analysis |
| LiveCodeBench | – | – | 71.5%#82 | – | Vals AI |
| Aider Polyglot | – | – | – | 60.4%#14 | Aider polyglot leaderboard |
| SWE-bench Verified (any scaffold)not in index | – | – | – | 42.4%#46 | SWE-bench |
Agents & tools
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| Cybench | – | 22.5%#10 | – | – | Cybench |
| Terminal-Bench Hard | – | – | 6.1%#257 | 6.8%#240 | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | – | – | 31.3%#247 | 28.6%#259 | Artificial Analysis |
Maths
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 3.9%#98 | 10.5%#89 | 18.6%#79 | – | Epoch AI Benchmarking Hub |
| FrontierMath Tier 4 | – | – | 0.0%#57 | – | Epoch AI Benchmarking Hub |
| OTIS Mock AIME | 44.4%#175 | 63.9%#140 | 76.9%#110 | – | Epoch AI Benchmarking Hub |
| MATH Level 5 | – | 95.2%#14 | 96.5%#10 | – | Epoch AI Benchmarking Hub |
| AIME (Vals) | – | – | 86.5%#32 | – | Vals AI |
| MGSM | – | – | 91.3%#27 | – | Vals AI |
Knowledge
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| SimpleQA Verified | – | – | 15.3%#67 | – | Epoch AI Benchmarking Hub |
| AA-Omniscience | – | – | -42.6#269 | – | Artificial Analysis |
| MMLU-Pro | – | – | 78.7%#97 | – | Vals AI |
| LegalBench | – | – | 71.5%#110 | – | Vals AI |
| CorpFin | – | – | 45.3%#111 | – | Vals AI |
| TaxEval | – | – | 69.4%#95 | – | Vals AI |
| MedQA | – | – | 94.8%#14 | – | Vals AI |
Instruction following
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| IFBench | – | – | 67.1%#88 | – | Artificial Analysis |
Human preference
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| LMArena Text | – | – | 1363#167 | 1348#181 | LMArena |
Long context
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| Fiction.LiveBench 120k | – | 43.8%#24 | – | – | Fiction.live |
| AA-LCR | – | – | 43.0%#277 | – | Artificial Analysis |
Composite
| Benchmark | low | medium | high | default | Source |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | 140.4#96 | Epoch AI Benchmarking Hub |
| AA Intelligence Index | – | – | 11.0#298 | 12.5#263 | Artificial Analysis |
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 | 109 tok/s | 5.49 s | $1.10 | $4.40 | 200k | – |
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 $0.550 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0018 | $0.0016 | 7.0 s |
| Summarise a 30-page report | 12,000 / 600 | $0.016 | $0.011 | 8.4 s |
| Code edit | 6,000 / 1,500 | $0.013 | $0.011 | 12.6 s |
| Agentic coding session | 60,000 / 4,000 | $0.084 | $0.059 | 24.4 s |
| Structured extraction | 2,000 / 200 | $0.0031 | $0.0023 | 6.5 s |
See also
Data as of 9 Sept 2026. Compare these configurations.