Hermes Best Models: leaderboard
Hermes agent-framework model evaluation: 25 agentic tasks (simple QA, multi-step tool use, delegation, reasoning, failure handling) run per LLM; task pass rate in percent, latency/cost excluded.
Metric: Pass Rate (%). Source: github.com. Saturation forecast: Estimated already saturated. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 100 | #255 |
| 2 | Gemini 3.1 Flash Lite | 100 | #200 |
| 3 | Claude Sonnet 4.6 | 96 | #94 |
| 4 | DeepSeek V4 Pro | 96 | #112 |
| 5 | GLM-5.1 | 96 | #109 |
| 6 | Claude Opus 4.6 | 92 | #67 |
| 7 | DeepSeek V3.2 | 92 | #210 |
| 8 | Kimi K2.6 | 92 | #113 |
| 9 | DeepSeek V4 Flash | 92 | #155 |
| 10 | Gemini 3 Flash (Preview) | 92 | #90 |
| 11 | Gemini 2.5 Flash Lite | 92 | #432 |
| 12 | MiniMax-M2.7 | 92 | #270 |
| 13 | Gemini 3.1 Pro (Preview) | 88 | #64 |
| 14 | GPT-5.5 | 88 | #33 |
| 15 | Claude Opus 4.7 | 88 | #51 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=hermes-best-models · How It Works · Data refreshed daily, snapshot 2026-11-05.