DeepSeekAuto-detectedDeepseek v4 Flash
Not yet ranked: too few independent results so far. Released 24 Apr 2026.
- Blended price
- –
- Output speed
- –
- First answer
- –
- Context
- –
- Reasoning63
- Coding54
- Human preference63
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 | – | – | – | – | – | Coding 56 |
| defaultbest | – | – | – | – | – | Coding 54 · Human preference 63 · Reasoning 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 | default | Source | Trend |
|---|---|---|---|---|
| LMArena Hard Prompts | – | 1458#83 | LMArena |
Coding
| Benchmark | high | default | Source | Trend |
|---|---|---|---|---|
| SciCode | 42.0%#99 | – | SciCode | |
| WeirdML | 43.8%#83 | – | WeirdML | |
| LMArena Coding | – | 1479#92 | LMArena | |
| LMArena WebDev | 1582#18 | 1431#63 | LMArena |
Human preference
| Benchmark | high | default | Source | Trend |
|---|---|---|---|---|
| LMArena Text | – | 1439#81 | LMArena |
See also
Data as of 9 Sept 2026. Compare these configurations.