BenchLM — leaderboard
Composite LLM leaderboard aggregating current model performance across agentic, coding, reasoning, grounded multimodal, knowledge, multilingual, instruction-following, and math categories.
Metric: Overall Score. Source: benchlm.ai. Status: saturation imminent. 200 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Mythos 5 | 83.9 |
| 2 | Claude Fable 5 | 83.7 |
| 3 | GPT-5.6 Sol | 82 |
| 4 | Kimi K3 | 81 |
| 5 | Claude Opus 4.8 | 78.3 |
| 6 | Muse Spark 1.1 | 77.4 |
| 7 | Grok 4.5 | 76.7 |
| 8 | GPT-5.4 | 74.2 |
| 9 | GPT-5.5 | 73.5 |
| 10 | Qwen 3.7 Max | 72.8 |
| 11 | GPT-5.6 Terra | 72.6 |
| 12 | Claude Opus 4.7 | 71.9 |
| 13 | Muse Spark | 71 |
| 14 | MiMo-V2.5-Pro | 70.2 |
| 15 | MiniMax-M3 | 69.8 |
Interactive version: theaggregate.ai/benchmark?slug=benchlm · How the rankings work · Data refreshed daily, snapshot 2026-07-22.