Head to head

GPT-OSS 20B vs Devstral Small 2

On a blended workload of one million input tokens and three million output tokens, GPT-OSS 20B is 1.5× cheaper than Devstral Small 2.

GPT-OSS 20BDevstral Small 2
Provider Together AI Mistral
Input / MTok $0.05 $0.1
Output / MTok $0.2 $0.3
Cache read / MTok
Context window 128,000
Blended 1M in + 3M out $0.65 $1.00
Status ga ga

GPT-OSS 20B detail   Devstral Small 2 detail