DeepSeek V4 Pro vs Kimi K3
DeepSeek V4 Pro 
Kimi K3 Verdict
- Kimi K3 leads on quality: 66.5 vs 59.5.
- DeepSeek V4 Pro (high) is stronger in instruction following.
- Kimi K3 is stronger in agents & tools, composite, knowledge, long context, coding, human preference, maths, multimodal.
- DeepSeek V4 Pro (high) is 11× cheaper ($0.548 vs $6.00 per 1M blended).
- DeepSeek V4 Pro (high) streams 1.7× faster (65 vs 38 tokens per second).
| Metric | DeepSeek V4 Pro (high) | Kimi K3 |
|---|---|---|
| BenchLeader Index | 59.5 | 66.5 |
| Agents & tools score | 63.8 | 67.4 |
| Composite score | 67.1 | 74.2 |
| Instruction following score | 69.6 | – |
| Knowledge score | 59.1 | 68.2 |
| Long context score | 61.6 | 71.1 |
| Coding score | – | 65.0 |
| Human preference score | – | 71.2 |
| Maths score | – | 78.1 |
| Multimodal score | – | 65.0 |
| Blended price $/M | $0.548 | $6.00 |
| Output speed | 65 tok/s | 38 tok/s |
| Time to first answer | 32.3 s | 56.3 s |
| Context window | 1M | 1.0M |
| SciCode | – | 58.7% |
| APEX-Agents | – | 39.3% |
| FrontierCode | – | 44.2% |
| ProofBench | – | 87.0% |
| Epoch Capabilities Index | – | 157.6 |
| LMArena Agent | 4.4 | – |
| LiveBench | – | 79.2% |
| LiveBench Reasoning | – | 90.7% |
| LiveBench Coding | – | 81.5% |
| LiveBench Agentic Coding | – | 62.2% |
| LiveBench Mathematics | – | 84.4% |
| LiveBench Data Analysis | – | 78.7% |
| LiveBench Language | – | 85.5% |
| AA Intelligence Index | 30.1 | 43.8 |
| IFBench | 71.3% | – |
| AA-LCR | 70.3% | 88.7% |
| MMMU-Pro | – | 80.5% |
| AA-Omniscience | -10.6 | 19.7 |
| Terminal-Bench Hard | 41.7% | – |
| GPQA Diamond (AA) | 90.5% | 93.5% |
| Humanity's Last Exam (AA) | 35.2% | 46.9% |
| SciCode (AA) | – | 59.5% |
| τ²-Bench Telecom (AA) | 94.2% | – |
| LegalBench | – | 86.0% |
| CorpFin | – | 71.6% |
| TaxEval | – | 75.7% |
| Terminal-Bench 2.1 (Vals) | – | 80.9% |
| SWE-bench (Vals) | – | 93.4% |
| GPQA Diamond (Vals) | – | 92.9% |
| Vals Index | – | 57.8 |
| EQ-Bench 4 | – | 1339 |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
DeepSeek V4 Pro vs Kimi K3: questions
- Is DeepSeek V4 Pro better than Kimi K3?
- Kimi K3 leads on quality: 66.5 vs 59.5. The BenchLeader Index combines every independent quality benchmark; Kimi K3 is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is DeepSeek V4 Pro better than Kimi K3 for agentic tasks?
- Kimi K3 scores higher in agentic tasks (67 vs 64 on the category index, where 50 is average).
- Which is cheaper, DeepSeek V4 Pro or Kimi K3?
- DeepSeek V4 Pro is cheaper: $0.548 against $6.00 per million tokens, blended at three input tokens per output token.
- Which is faster, DeepSeek V4 Pro or Kimi K3?
- DeepSeek V4 Pro streams faster: 65 against 38 output tokens per second.
- Which has the larger context window?
- Kimi K3 accepts more context: 1.0M against 1M tokens.