Kimi K3 scores 50.6 on the Respan Index (#21), ahead of Grok 4.5 at 43.2 (#38). Kimi K3 leads on 19 of the 25 benchmarks both report. Grok 4.5 is 2.0x cheaper per token at list price. Kimi K3 has the larger context window (1.0M tokens).
* Estimated: no published scores in that category. How the Respan Index works
| 57.2% |
| 72.2% |
| Furniture Assembly (Epoch AI) | 22.5% | 34.2% |
| GPQA Diamond | 93.4% | 93.5% |
| LiveBench Agentic Coding (LiveBench) | 56.5% | 62.2% |
| LiveBench Coding (LiveBench) | 68.6% | 81.4% |
| LiveBench Data Analysis (LiveBench) | 73% | 78.7% |
| LiveBench Instruction Following (LiveBench) | 71.5% | 71.4% |
| LiveBench Language (LiveBench) | 82.8% | 85.5% |
| LiveBench Mathematics (LiveBench) | 90.8% | 84.4% |
| LiveBench Reasoning (LiveBench) | 87.2% | 90.7% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 97.8% | 97.2% |
| SAGE (Vals AI) | 35% | 52.8% |
| SimpleQA Verified | 48.3% | 50.6% |
| Terminal-Bench 4.0 (Vals AI) | 8.6% | 17.2% |
| Vending-Bench 2 (Andon Labs) | 3887.43 | 5165.04 |
| WeirdML (Håvard Tveit Ihle) | 46.4% | 82.6% |
| LMArena Agent (LMArena) | 0.0121 | 0.0418 |
| LMArena Elo | 1450 | 1488 |
| LMArena WebDev (LMArena) | 1552 | 1658 |
| tau2-bench Banking Knowledge (Sierra) | 47.9% | 37.1% |
| SimpleBench (SimpleBench) | 70% | 60.7% |
Higher is better. Each score links to where it was published.