OpenAIReasoning modelGPT-5.1-Codex
Best configuration ranks #149 of 610 on the BenchLeader Index at 57.1 ±5.2. Last measured 8 Sept 2026. Released 14 Nov 2025.
- Blended price
- $3.44/M
- $1.25 in · $10.00 out
- Output speed
- 48 tok/s
- First answer
- 5.18 s
- Context
- 400k
- Overall index57
- Coding50
- Agents & tools58
- Knowledge61
- Instruction following68
- Multimodal57
- Long context61
- Composite59
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 |
|---|---|---|---|---|---|---|
| medium | – | – | 48 tok/s | 5.18 s | $0.0035 | Coding 58 |
| high | – | – | 48 tok/s | 5.18 s | $0.0035 | Coding 38 |
| defaultbest | 57.1 | #149 | 48 tok/s | 5.18 s | $0.0035 | Agents & tools 58 · Coding 50 · Composite 59 · Instruction following 68 · Knowledge 61 · Long context 61 · Multimodal 57 |
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 | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamond (AA)not in index | – | – | 86.0%#115 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | – | 25.7%#132 | Artificial Analysis |
Coding
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| LMArena WebDev | – | – | 1336#95 | LMArena | |
| LiveCodeBench | – | – | 85.5%#30 | Vals AI | |
| IOI | – | 3.7%#50 | – | Vals AI | |
| SWE-bench Verified (bash only) | 66.0%#17 | – | – | SWE-bench | |
| SWE-bench Verified (any scaffold)not in index | 66.0%#21 | – | – | SWE-bench |
Agents & tools
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| Terminal-Bench | – | – | 57.8%#19 | Terminal-Bench | |
| APEX-Agents | – | – | 20.7%#35 | Mercor | |
| Terminal-Bench Hard | – | – | 34.9%#80 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | – | – | 83.0%#109 | Artificial Analysis |
Knowledge
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| AA-Omniscience | – | – | -6.5#118 | Artificial Analysis |
Instruction following
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| IFBench | – | – | 70.0%#70 | Artificial Analysis |
Multimodal
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| MMMU-Pro | – | – | 72.5%#103 | Artificial Analysis |
Long context
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| AA-LCR | – | – | 69.3%#170 | Artificial Analysis |
Composite
| Benchmark | medium | high | default | Source | Trend |
|---|---|---|---|---|---|
| AA Intelligence Index | – | – | 23.7#127 | 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 |
|---|---|---|---|---|---|---|
| Azure | 48 tok/s | 5.18 s | $1.25 | $10.00 | 400k | – |
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 $0.125 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0035 | $0.0032 | 11.4 s |
| Summarise a 30-page report | 12,000 / 600 | $0.021 | $0.011 | 17.7 s |
| Code edit | 6,000 / 1,500 | $0.022 | $0.017 | 36.4 s |
| Agentic coding session | 60,000 / 4,000 | $0.115 | $0.064 | 1.5 min |
| Structured extraction | 2,000 / 200 | $0.0045 | $0.0028 | 9.4 s |
See also
Data as of 9 Sept 2026. Compare these configurations.