GLM 4.7 vs Kimi K2 Thinking
Côte à côte. Chaque métrique. Chaque benchmark.
| Type | GLM 4.7 | Kimi K2 Thinking |
|---|---|---|
| Provider |  z-ai |  moonshotai |
| score moyen | 50.5 | 53.3 |
| Prix d'entrée | $0.39 | $0.60 |
| Prix de sortie | $1.75 | $2.50 |
| Fenêtre de contexte | 203K tokens (~101 books) | 262K tokens (~131 books) |
| Sorti le | 2025-12-22 | 2025-11-06 |
| Code source ouvert | Open Source | Open Source |
Scores de benchmark
23 benchmarks · GLM 4.7: 9, Kimi K2 Thinking: 13
| Benchmark | Catégorie | GLM 4.7 | Kimi K2 Thinking |
|---|---|---|---|
| APEX-Agents | agentic | 3.1 | 4.0 |
| Chess Puzzles | knowledge | 6.0 | 20.0 |
| FrontierMath-2025-02-28-Private | math | 2.4 | 21.4 |
| FrontierMath-Tier-4-2025-07-01-Private | math | 0.1 | 0.1 |
| GPQA diamond | knowledge | 77.8 | 79.0 |
| LiveBench — Agentic Coding | coding | 41.7 | 38.3 |
| LiveBench — Coding | coding | 73.1 | 67.4 |
| LiveBench — Data Analysis | reasoning | 55.2 | 52.3 |
| LiveBench — If | language | 35.7 | 62.0 |
| LiveBench — Language | language | 65.2 | 66.5 |
| LiveBench — Mathematics | math | 76.0 | 81.1 |
| LiveBench — Overall | knowledge | 58.1 | 61.6 |
| LiveBench — Reasoning | reasoning | 59.7 | 63.5 |
| OpenCompass — AIME2025 | math | 95.4 | 94.1 |
| OpenCompass — GPQA-Diamond | knowledge | 86.9 | 82.7 |
| OpenCompass — HLE | knowledge | 25.4 | 21.3 |
| OpenCompass — IFEval | language | 90.2 | 92.4 |
| OpenCompass — LiveCodeBenchV6 | coding | 83.8 | 77.1 |
| OpenCompass — MMLU-Pro | knowledge | 84.0 | 84.3 |
| OTIS Mock AIME 2024-2025 | math | 83.3 | 83.0 |
| PostTrainBench | knowledge | 7.5 | 7.3 |
| SimpleQA Verified | knowledge | 31.5 | 31.6 |
| Terminal Bench | coding | 33.4 | 35.7 |