GLM 5.1 vs Kimi K3
GLM 5.1 
Kimi K3 Verdict
- Kimi K3 leads on quality: 66.5 vs 60.2.
- GLM 5.1 is stronger in instruction following, reasoning.
- Kimi K3 is stronger in agents & tools, coding, composite, human preference, knowledge, long context, maths, multimodal.
- GLM 5.1 is 2.8× cheaper ($2.15 vs $6.00 per 1M blended).
- GLM 5.1 streams 1.7× faster (65 vs 38 tokens per second).
| Metric | GLM 5.1 | Kimi K3 |
|---|---|---|
| BenchLeader Index | 60.2 | 66.5 |
| Agents & tools score | 56.8 | 67.4 |
| Coding score | 57.3 | 65.0 |
| Composite score | 62.4 | 74.2 |
| Human preference score | 66.6 | 71.2 |
| Instruction following score | 73.9 | – |
| Knowledge score | 56.6 | 68.2 |
| Long context score | 63.3 | 71.1 |
| Maths score | 53.0 | 78.1 |
| Reasoning score | 61.4 | – |
| Multimodal score | – | 65.0 |
| Blended price $/M | $2.15 | $6.00 |
| Output speed | 65 tok/s | 38 tok/s |
| Time to first answer | 60.1 s | 56.3 s |
| Context window | 200k | 1.0M |
| GPQA Diamond | 89.9% | – |
| FrontierMath Tiers 1–3 | 36.8% | – |
| OTIS Mock AIME | 93.3% | – |
| SWE-bench Verified (Epoch) | 74.2% | – |
| SimpleQA Verified | 34.0% | – |
| SimpleBench | 55.1% | – |
| SciCode | 43.8% | 58.7% |
| WeirdML | 57.1% | – |
| APEX-Agents | – | 39.3% |
| FrontierCode | – | 44.2% |
| ProofBench | 22.2% | 87.0% |
| Epoch Capabilities Index | 149.7 | 157.6 |
| LMArena Text | 1466 | – |
| LMArena Hard Prompts | 1489 | – |
| LMArena Coding | 1514 | – |
| LMArena WebDev | 1508 | – |
| 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 | 26.4 | 43.8 |
| IFBench | 76.3% | – |
| AA-LCR | 73.7% | 88.7% |
| MMMU-Pro | – | 80.5% |
| AA-Omniscience | 0.8 | 19.7 |
| Terminal-Bench Hard | 43.2% | – |
| GPQA Diamond (AA) | 86.8% | 93.5% |
| Humanity's Last Exam (AA) | 30.1% | 46.9% |
| SciCode (AA) | 44.8% | 59.5% |
| τ²-Bench Telecom (AA) | 97.7% | – |
| AIME (Vals) | 91.9% | – |
| LiveCodeBench | 81.4% | – |
| MMLU-Pro | 86.9% | – |
| LegalBench | 84.4% | 86.0% |
| CorpFin | 64.5% | 71.6% |
| TaxEval | 71.2% | 75.7% |
| Terminal-Bench 2.1 (Vals) | 56.9% | 80.9% |
| SWE-bench (Vals) | 76.4% | 93.4% |
| GPQA Diamond (Vals) | 84.5% | 92.9% |
| Vals Index | – | 57.8 |
| AIME 2026 | 95.8% | – |
| HMMT February 2026 | 89.4% | – |
| MathArena Apex | 11.5% | – |
| MCP Atlas | 75.6% | – |
| 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.
GLM 5.1 vs Kimi K3: questions
- Is GLM 5.1 better than Kimi K3?
- Kimi K3 leads on quality: 66.5 vs 60.2. 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 GLM 5.1 better than Kimi K3 for coding?
- Kimi K3 scores higher in coding (65 vs 57 on the category index, where 50 is average).
- Is GLM 5.1 better than Kimi K3 for agentic tasks?
- Kimi K3 scores higher in agentic tasks (67 vs 57 on the category index, where 50 is average).
- Which is cheaper, GLM 5.1 or Kimi K3?
- GLM 5.1 is cheaper: $2.15 against $6.00 per million tokens, blended at three input tokens per output token.
- Which is faster, GLM 5.1 or Kimi K3?
- GLM 5.1 streams faster: 65 against 38 output tokens per second.
- Which has the larger context window?
- Kimi K3 accepts more context: 1.0M against 200k tokens.