DeepSeek V4 Flash scores 43.3 on the Respan Index (#37), ahead of Kimi K2.6 at 34.4 (#55). DeepSeek V4 Flash leads on 12 of the 18 benchmarks both report. DeepSeek V4 Flash has the larger context window (1M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 46.8% |
| 46.9% |
| LiveBench Coding (LiveBench) | 75% | 78.6% |
| LiveBench Data Analysis (LiveBench) | 79.3% | 65.1% |
| LiveBench Instruction Following (LiveBench) | 65.5% | 64.4% |
| LiveBench Language (LiveBench) | 79.2% | 75.1% |
| LiveBench Mathematics (LiveBench) | 86.8% | 84.3% |
| LiveBench Reasoning (LiveBench) | 86.6% | 79.4% |
| Mystery Game Puzzles (Epoch AI) | 34% | 18% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 94.4% | 96.1% |
| SimpleQA Verified | 33.6% | 34.9% |
| WeirdML (Håvard Tveit Ihle) | 63% | 55.9% |
| LMArena Elo | 1436 | 1455 |
| LMArena WebDev (LMArena) | 1581 | 1509 |
| Toolathlon-Verified (HKUST) | 70.7% | 58% |
Higher is better. Each score links to where it was published.