DeepSeekAuto-detectedDeepseek v4 Pro
Best configuration ranks #91 of 610 on the BenchLeader Index at 60.3 ±4.7 (high reasoning effort). Last measured 8 Sept 2026. Released 24 Apr 2026.
- Blended price
- –
- Output speed
- –
- First answer
- –
- Context
- –
- Overall index60
- Reasoning66
- Coding59
- Maths66
- Human preference66
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 |
|---|---|---|---|---|---|---|
| highbest | 60.3 | #91 | – | – | – | Coding 59 · Human preference 66 · Maths 66 · Reasoning 66 |
| default | – | – | – | – | – | Coding 56 · Human preference 65 · Reasoning 65 |
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 | default | Source | Trend |
|---|---|---|---|---|
| GPQA Diamond | 90.9%#29 | – | Epoch AI Benchmarking Hub | |
| LMArena Hard Prompts | 1482#51 | 1474#64 | LMArena |
Coding
| Benchmark | high | default | Source | Trend |
|---|---|---|---|---|
| SciCode | 46.4%#80 | – | SciCode | |
| WeirdML | 46.5%#75 | – | WeirdML | |
| LMArena Coding | 1505#55 | 1489#80 | LMArena | |
| LMArena WebDev | 1580#19 | 1464#49 | LMArena |
Maths
| Benchmark | high | default | Source | Trend |
|---|---|---|---|---|
| OTIS Mock AIME | 95.6%#29 | – | Epoch AI Benchmarking Hub |
Human preference
| Benchmark | high | default | Source | Trend |
|---|---|---|---|---|
| LMArena Text | 1460#51 | 1455#60 | LMArena |
See also
Data as of 9 Sept 2026. Compare these configurations.