DeepSeek V3
Change Analysis · 2026 Week15
DeepSeek V3 2026 Week15 Code Execution (v5) dimension dropped 8.4 pts
Score Comparison
75.1
72.7
-2.4
| Dimension | Previous | Current | Change |
|---|---|---|---|
| Code Execution (v5) | 92.2 | 83.8 | -8.4 |
| Knowledge Synthesis (v5) | 47.6 | 48.3 | +0.7 |
| Grounding (v5) | 78.8 | 77.2 | -1.6 |
| Value | 99.7 | 99.6 | -0.1 |
| Stability | 33.6 | 31.3 | -2.3 |
| Availability | 100 | 100 | 0 |
All matched tasks had no score changes, or no tasks could be matched to the previous evaluation.
Back to Movers