Mistral Small 3.2 24 B Instruct vs Llama 4 Scout: cost comparison
Prices updated 2026-11-11 · computed from live provider data
Short answer: Llama 4 Scout is cheaper on a
3:1 input/output blend — $0.150 per million tokens versus
$0.150, a 0% difference.
Llama 4 Scout offers the larger context window at 1,310,720 tokens, against 256,000 for Mistral Small 3.2 24 B Instruct.
Head to head
| Mistral Small 3.2 24 B Instruct | Llama 4 Scout | |
|---|---|---|
| Input, per 1M tokens | $0.100 | $0.100 |
| Output, per 1M tokens | $0.300 | $0.300 |
| Blended 3:1, per 1M | $0.150 | $0.150 |
| Context window | 256,000 | 1,310,720 |
| Vendor | mistralai | meta-llama |
Monthly cost at real volumes
| Workload (input / output) | Mistral Small 3.2 24 B Instruct | Llama 4 Scout |
|---|---|---|
| 1M in / 250K out | $0.175 | $0.175 |
| 10M in / 2.5M out | $1.75 | $1.75 |
| 100M in / 25M out | $17.50 | $17.50 |
Which should you pick?
Pick Llama 4 Scout when you need the longer context window (1,310,720 vs 256,000 tokens). Pick Llama 4 Scout when price per token dominates your workload — high-volume classification, extraction, or batch summarisation.