models / devstral-2
Devstral 2
mistralai/Devstral-2-123B-Instruct-2512params
125B
sparsity
Dense
attention
Full attention
architecture
ministral3
license
other
downloads
45.0k
likes
327
hf updated
Jun 30, 2026
Visualize
total size
each column is a layer — every parameter fires on every token
bits / param
16-bit
footprint
233 GiB
sparks to fit
3×
block ≈
977M params
active paramsghost = precision traded awayblock fill ∝ bits kept · Devstral 2 ≈ 125.02598984B params
Eval scores (compare all)
Reported
72.2SWE-bench Verified
Benchmarks (0)
No benchmarks for this model yet.