Nous Research · Hermes 4
Hermes 4 405B
Nous Research's open-weights hybrid reasoning fine-tune of Llama 3.1 405B.
- AA Intelligence
- 7.5#119 of 142
- Input / output per 1M
- $1.00/ $3.00
- Output speed
- 42tok/s
- Context window
- 128K tokens
Composite indexes
- Artificial Analysis Intelligence Index#119 of 1428
Leader: Claude Opus 5.5 at 58
Coding
- LiveCodeBench#28 of 5668.6%
Leader: Fugu at 92.9%
Agents and tool use
- τ²-Bench#64 of 8822.2%
Leader: GLM-5.2 at 99.1%
Reasoning and knowledge
- GPQA Diamond#96 of 14272.7%
Leader: GPT-6 Astra at 96.1%
- Humanity’s Last Exam#98 of 14710.9%
Leader: Claude Opus 5.5 at 61.4%
- CritPt#89 of 1390.3%
Leader: GPT-5.6 Sol at 32.3%
Vision and long context
- AA Long Context Reasoning#112 of 13922.3%
Leader: Kimi K3 at 88.7%
Other published results
| Result | Score | Reported by | Source |
|---|---|---|---|
| aime-2025 · AIME 2025; reasoning | 69.7% | independent | Artificial Analysis: Hermes 4 - Llama-3.1 405B (Reasoning) ↗ |
| Terminal-Bench Hard (AA) · reasoning | 11.4% | independent | Artificial Analysis: Hermes 4 - Llama-3.1 405B (Reasoning) ↗ |
| IFBench · reasoning | 32.7% | independent | Artificial Analysis: Hermes 4 - Llama-3.1 405B (Reasoning) ↗ |
| AA-Omniscience Index · reasoning | -36 index | independent | Artificial Analysis: Hermes 4 - Llama-3.1 405B (Reasoning) ↗ |