DeepSeek V4.1 Flash vs Kimi K3
DeepSeek V4.1 Flash 
Kimi K3 Verdict
- Kimi K3 (max) leads on quality: 67.2 vs 61.9.
- DeepSeek V4.1 Flash (max) is stronger in coding, reasoning.
- Kimi K3 (max) is stronger in agents & tools, composite, knowledge, long context, multimodal, human preference, maths.
- DeepSeek V4.1 Flash (max) is 23× cheaper ($0.262 vs $6.00 per 1M blended).
- DeepSeek V4.1 Flash (max) streams 5.8× faster (215 vs 37 tokens per second).
| Metric | DeepSeek V4.1 Flash (max) | Kimi K3 (max) |
|---|---|---|
| BenchLeader Index | 61.9 | 67.2 |
| Agents & tools score | 59.6 | 67.5 |
| Coding score | 68.9 | 67.0 |
| Composite score | 74.5 | 84.5 |
| Knowledge score | 61.7 | 64.5 |
| Long context score | 68.5 | 71.0 |
| Multimodal score | 61.4 | 65.1 |
| Reasoning score | 69.5 | 68.6 |
| Human preference score | – | 68.8 |
| Maths score | – | 65.4 |
| Blended price $/M | $0.262 | $6.00 |
| Output speed | 215 tok/s | 37 tok/s |
| Time to first answer | 10.5 s | 58.3 s |
| Context window | 1M | 1.0M |
| GPQA Diamond | – | 93.1% |
| FrontierMath Tiers 1–3 | – | 72.2% |
| FrontierMath Tier 4 | – | 39.0% |
| OTIS Mock AIME | – | 97.2% |
| SimpleQA Verified | – | 50.6% |
| SimpleBench | – | 60.7% |
| SciCode | – | 58.7% |
| WeirdML | – | 82.6% |
| LMArena Text | – | 1485 |
| LMArena Hard Prompts | – | 1514 |
| LMArena Coding | – | 1538 |
| LMArena WebDev | 1614 | 1674 |
| LMArena Agent | 4.9 | 6.2 |
| LiveBench | 81.1% | – |
| LiveBench Reasoning | 86.7% | – |
| LiveBench Coding | 80.0% | – |
| LiveBench Agentic Coding | 77.3% | – |
| LiveBench Mathematics | 93.3% | – |
| LiveBench Data Analysis | 79.3% | – |
| LiveBench Language | 81.2% | – |
| LiveBench Instruction Following | 70.0% | – |
| AA Intelligence Index | 39.5 | 43.8 |
| AA-LCR | 84.0% | 88.7% |
| MMMU-Pro | 77.0% | 80.5% |
| AA-Omniscience | -5.3 | 19.7 |
| GPQA Diamond (AA) | – | 93.5% |
| Humanity's Last Exam (AA) | 39.3% | 46.9% |
| SciCode (AA) | 51.9% | 59.5% |
| LiveCodeBench | – | 87.2% |
| MMLU-Pro | – | 88.0% |
| IOI | – | 48.9% |
| MCP Atlas | – | 82.3% |
| ARC-AGI-1 | – | 94.5% |
| ARC-AGI-2 | – | 60.4% |
| CritPt | 14.3% | 23.4% |
| GDPval (AA) | 56.6% | 52.4% |
| τ²-Bench Banking (AA) | – | 46.0% |
| ITBench SRE (AA) | – | 47.7% |
| Analyst Agent (AA) | – | 38.8% |
| APEX-Agents (AA) | – | 41.3% |
| BioMysteryBench | – | 71.5% |
| Code Migration | – | 16.1% |
| Excel Modeling Benchmark | – | 66.4% |
| Tax Agent Bench | – | 68.7% |
| Terminal-Bench 4.0 (Vals) | – | 12.6% |
| Terminal-Bench Science | – | 2.9% |
| Time Horizon Index: KSP | – | 10.5% |
| Vibe Code Bench 1-100 | – | 18.2% |
| FORTRESS | – | 26.6% |
| LMArena Maths | – | 1501 |
| LMArena Creative Writing | – | 1458 |
| LMArena Instruction Following | – | 1484 |
| LMArena Multi-turn | – | 1496 |
| LMArena Longer Queries | – | 1500 |
| Chess Puzzles | – | 39.0% |
| Mystery Game Puzzles | – | 26.0% |
| Surface Evolver Bench | – | 93.0% |
| DeepSWE | – | 68.5% |
| LMCA | – | 52.7% |
| DTBench | – | 91.2% |
| CursorBench | – | 60.8% |
| ALE-Bench | – | 1524.5 |
| GDP.pdf | – | 19.0% |
| FrontierSWE | – | 25.9% |
Data as of 2026-09-19. Best configuration of each model; every score links to its source on the model pages.
DeepSeek V4.1 Flash vs Kimi K3: questions
- Is DeepSeek V4.1 Flash better than Kimi K3?
- Kimi K3 (max) leads on quality: 67.2 vs 61.9. The BenchLeader Index combines every independent quality benchmark; Kimi K3 (max) is ahead overall as of 2026-09-19, but check the category scores for your use.
- Is DeepSeek V4.1 Flash better than Kimi K3 for coding?
- DeepSeek V4.1 Flash scores higher in coding (69 vs 67 on the category index, where 50 is average).
- Is DeepSeek V4.1 Flash better than Kimi K3 for agentic tasks?
- Kimi K3 scores higher in agentic tasks (68 vs 60 on the category index, where 50 is average).
- Which is cheaper, DeepSeek V4.1 Flash or Kimi K3?
- DeepSeek V4.1 Flash is cheaper: $0.262 against $6.00 per million tokens, blended at three input tokens per output token.
- Which is faster, DeepSeek V4.1 Flash or Kimi K3?
- DeepSeek V4.1 Flash streams faster: 215 against 37 output tokens per second.
- Which has the larger context window?
- Kimi K3 accepts more context: 1.0M against 1M tokens.