Llama 3.2 90B Instruct vs Sonar Reasoning Pro
Meta vs Perplexity — benchmarks, pricing, and capabilities side by side.
- •Sonar Reasoning Pro has the higher intelligence index (95.7 vs 54)
- •Llama 3.2 90B Instruct is cheaper ($0.35 vs $2.00 per 1M input)
| Llama 3.2 90B Instruct | Sonar Reasoning Pro | |
|---|---|---|
| Intelligence index | 54 | 95.7 |
| Developer | Meta | Perplexity |
| Type | Multimodal | Multimodal |
| Access | Open weights | API only |
| Context window | 128,000 tokens | 128,000 tokens |
| Input price | $0.35 / 1M | $2.00 / 1M |
| Output price | $0.40 / 1M | $8.00 / 1M |
| Speed | 100 tok/s | — |
| Released | September 25, 2024 | March 7, 2025 |
| Parameters | 90000000000 | — |
| Input modalities | — | Text, Image |
| Output modalities | — | Text |
Shared benchmarks
Llama 3.2 90B Instruct
Sonar Reasoning Pro
MATH-500
62.9
95.7