Leaderboard / Nous Research (Hermes)
Hermes 4.3 36B
Nous Research (Hermes) · Hermes 4 · released 2025-11-17 · data checked 2026-11-01
Hermes 4.3 36B is a 36.2B-parameter dense model from Nous Research (Hermes) (United States), released 2025-11-17 under the Apache-2.0 license. It scores 51.4 on the Small LM Index, #27 of 34 medium models. It needs about 25.3 GB of memory at 4-bit with an 8K-token context.
Post-trained on the decentralised Psyche network.
Benchmarks
Each bar shows this model's score. The tick marks the best medium model on that benchmark. Sources: 6 reported by Nous Research (Hermes). How we label sources.
Not reported: estimated
What the index model expects from this model's other scores. Not a measurement.
Specifications
- Lab
- Nous Research (Hermes) (US · United States)
- Based on
- Seed-OSS 36B (fine-tune or distillation)
- Architecture
- Dense transformer
- Total parameters
- 36.2B
- Active parameters
- All (dense)
- Size class
- Medium
- Context window
- 512K tokens
- Input modalities
- text
- Reasoning
- Hybrid (thinking on/off)
- License
- Apache-2.0
- Openness
- Open weights
- Memory to run
- ~25.3 GB at 4-bit, 8K context (more sizes)
- Hugging Face
- 2.2K downloads (30 days) · 312 likes
Compare with
- Hermes 4.3 36B vs Seed-OSS 36B Instruct (36.2B, index 60.3)
- Hermes 4.3 36B vs MiniCPM-V 4.7 35B-A3B (35.2B, index 54.5)
- Hermes 4.3 36B vs Qwen3.6 35B-A3B (35B, index 72.2)
- Hermes 4.3 36B vs Qwen3.5 35B-A3B (35B, index 69.9)
Sources
Something wrong here? Report it.
Run it locally
No GGUF build from a well-known publisher yet.
vllm serve NousResearch/Hermes-4.3-36BMemory needed: weights, context cache and runtime buffers.
| Quantization | 4K context | 32K context | 256K context |
|---|---|---|---|
| 16-bit F16 | 75 GB | 82 GB | 142 GB |
| 8-bit Q8_0 | 41 GB | 48 GB | 108 GB |
| 4-bit Q4_K_M | 24 GB | 32 GB | 92 GB |