"Mistral and Inkling do not do very well (both are run in opencode)" (WeirdML v3)
CodeRead the post
x.com · link re-checked 2026-10-16
Mistral Medium 3.5
By Mistral. 3 results on 3 boards, copied as published. Different boards, different metrics: don't add them up.
License (as listed)Modified MIT
Hosting regionNot confirmed
Boards3
Code
- 66.4%Vals AI — SWE-bench Verified
Mistral Medium 3.5 · % resolved · board #72
Mini-SWE-agent · as of 2026-09-01 Stale ↗
Agentic / terminal
No result on the boards we captured
Websites
Design
No result on the boards we captured
Android
No result on the boards we captured
iOS
No mature public leaderboard
Available in
No plan row in our pricing data names this model version.
Use cases with this model
No sourced use case names this exact model version.
Same family: Mistral/Devstral 10
"Mistral Large 4 scores 13.3%" on sudo L7 staff-level SWE benchmark
CodeRead the post
x.com · link re-checked 2026-10-16
higher reasoning effort scored worse and cost 8.3x more
CodeRead the post
x.com · link re-checked 2026-10-16
"qwen 3.8 flash next … built the whole thing on my dgx spark in 2.5 hours", where Mistral Large 4 "burned 233k tokens and never wrote a single file"
CodeRead the post
x.com · link re-checked 2026-10-16
Cost (user-reported): "Mistral large 4 costs $14 vs GPT 6.1 $1" for the same revenue-forecast coding job
CodeRead the post
x.com · link re-checked 2026-10-16
"Mistral Medium isn't that great for coding because of its small context window of 200K"
CodeRead the post
x.com · link re-checked 2026-10-16
All 10 Mistral/Devstral use cases →
Sources
lmarena.ai/leaderboard/webdev vals.ai/benchmarks/vibe-code vals.ai/benchmarks/swebench