Hermes 3 - Llama-3.1 70B vs Sonar Reasoning Pro
Nous Research vs Perplexity — benchmarks, pricing, and capabilities side by side.
- •Sonar Reasoning Pro has the higher intelligence index (95.7 vs 42.5)
- •Hermes 3 - Llama-3.1 70B is cheaper ($0.30 vs $2.00 per 1M input)
| Hermes 3 - Llama-3.1 70B | Sonar Reasoning Pro | |
|---|---|---|
| Intelligence index | 42.5 | 95.7 |
| Developer | Nous Research | Perplexity |
| Type | LLM | Multimodal |
| Access | — | API only |
| Context window | — | 128,000 tokens |
| Input price | $0.30 / 1M | $2.00 / 1M |
| Output price | $0.30 / 1M | $8.00 / 1M |
| Speed | 32 tok/s | — |
| Released | August 15, 2024 | March 7, 2025 |
| Parameters | — | — |
| Input modalities | — | Text, Image |
| Output modalities | — | Text |
Shared benchmarks
Hermes 3 - Llama-3.1 70B
Sonar Reasoning Pro
MATH-500
53.8
95.7