OpenAIGPT-5.3 Codex
Best configuration ranks #54 of 610 on the BenchLeader Index at 62.4 ±9.0. Last measured 8 Sept 2026. Released 5 Feb 2026.
- Blended price
- $4.81/M
- $1.75 in · $14.00 out
- Output speed
- 126 tok/s
- First answer
- 65 s
- first token 3.89 s
- Context
- 400k
- Overall index62
- Coding55
- Agents & tools63
- Knowledge69
- Instruction following73
- Multimodal63
- Long context68
- Composite70
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 |
|---|---|---|---|---|---|---|
| high | – | – | 126 tok/s | 65 s | $0.0049 | Coding 56 |
| xhigh | – | – | 126 tok/s | 65 s | $0.0049 | Coding 65 |
| defaultbest | 62.4 | #54 | 126 tok/s | 65 s | $0.0049 | Agents & tools 63 · Coding 55 · Composite 70 · Instruction following 73 · Knowledge 69 · Long context 68 · Multimodal 63 |
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 | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| GPQA Diamond (AA)not in index | – | – | 91.5%#46 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | – | 42.5%#38 | Artificial Analysis |
Coding
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | 74.8%#15 | – | – | Epoch AI Benchmarking Hub | |
| WeirdML | – | 77.9%#19 | 79.3%#16 | WeirdML | |
| LMArena WebDev | – | – | 1409#66 | LMArena | |
| LiveCodeBench | – | 87.3%#15 | – | Vals AI | |
| IOI | – | 43.8%#12 | – | Vals AI | |
| SWE-bench (Vals)not in index | – | 78.0%#33 | – | Vals AI |
Agents & tools
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| Terminal-Bench | – | – | 78.4%#6 | Terminal-Bench | |
| APEX-Agents | – | – | 31.8%#26 | Mercor | |
| Terminal-Bench Hard | – | – | 53.0%#17 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | – | – | 86.0%#91 | Artificial Analysis | |
| HiL-Bench | – | – | 4.3%#17 | Scale AI SEAL |
Knowledge
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| AA-Omniscience | – | – | 10.9#62 | Artificial Analysis |
Instruction following
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| IFBench | – | – | 75.4%#33 | Artificial Analysis |
Multimodal
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| MMMU-Pro | – | – | 78.5%#52 | Artificial Analysis |
Long context
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| AA-LCR | – | – | 83.3%#10 | Artificial Analysis |
Composite
| Benchmark | high | xhigh | default | Source | Trend |
|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | 156.6#15 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index | – | – | 32.5#70 | 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 | 63 tok/s | 2.38 s | $1.75 | $14.00 | 400k | – |
| Azure | 50 tok/s | 5.40 s | $1.75 | $14.00 | 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 $0.175 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0049 | $0.0044 | 1.1 min |
| Summarise a 30-page report | 12,000 / 600 | $0.029 | $0.015 | 1.2 min |
| Code edit | 6,000 / 1,500 | $0.032 | $0.024 | 1.3 min |
| Agentic coding session | 60,000 / 4,000 | $0.161 | $0.090 | 1.6 min |
| Structured extraction | 2,000 / 200 | $0.0063 | $0.0039 | 1.1 min |
See also
Data as of 9 Sept 2026. Compare these configurations.