Devstral

model
Mistral
Agentic coding model family. Devstral 2 (123B, Dec 2025) achieves 72.2% SWE-Bench Verified. Devstral Small 2 (24B) achieves 68.0%. Runs on single RTX 4090. Collaboration with All Hands AI.
Paper (arXiv)OpenRouter

Outputs 2

Devstral (24B)

model
Paper (arXiv)
Architecture DENSE
Parameters 24B
Context window 128,000

arXiv: 2509.25193

Devstral 2 (123B)

model

72.2% SWE-Bench Verified. 256K context.

OpenRouter
Architecture DENSE
Parameters 123B
Context window 256,000
codingagenticopen-weight