Up to four models
DeepSeek-V3.1 vs DeepSeek-V3 (Mar 2025)
On the index (GPT-4 = 100), DeepSeek-V3.1 scores 128 and DeepSeek-V3 (Mar 2025) 124. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
128111–145
WeightsOpen
ReleasedAug 2025
Price in / out$0.25 / $0.95
Context164K
Core tests taken0 of 9
124108–140
WeightsOpen
ReleasedMar 2025
Price in / out$0.29 / $1.14
Context164K
Core tests taken2 of 9
DeepSeek-V3.1 and DeepSeek-V3 (Mar 2025) are too close to call (4 points apart, within the uncertainty), at a similar price.
Test by testthird-party results
Every result as a table
| Test | DeepSeek-V3.1 | DeepSeek-V3 (Mar 2025) |
|---|---|---|
| Reasoning & knowledge | ||
| Fiction.LiveBench | 52.8% | 50.0% |
| SimpleBench | 40.0% | 27.2% |
| Coding | ||
| WeirdML | 38.4% | 36.1% |
| Agents | ||
| DTBench | 82.7% | 64.8% |
| LMCA | 24.3% | 15.5% |
Task length · METR
30 sec
Tests at least two of these models have taken. Bold rows are the core tests.