Devstral Small (Jul '25) vs Mistral Medium 3.5

Wins 0 of 5 benchmarks
$0.50 / task
Wins 0 of 5 benchmarks
Different efforts: non-thinking · high
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Devstral Small (Jul '25)—
Mistral Medium 3.513.8
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Devstral Small (Jul '25)—
Mistral Medium 3.50
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Devstral Small (Jul '25)—
Mistral Medium 3.56.3
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Devstral Small (Jul '25)—
Mistral Medium 3.51,400
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Devstral Small (Jul '25)—
Mistral Medium 3.564.9

Specs

SpecDevstral Small (Jul '25)Mistral Medium 3.5
Input, per 1M$0.10$1.50
Output, per 1M$0.30$7.50
Cost per task?—$0.50
Speed, tok/s—155
Context256K262K
WeightsOpenOpen
ReleasedJul 2025Apr 2026
Memory (4-bit)14 GB77 GB
LicenceApache-2.0custom licence
Commercial useAllowedRestricted
See images—✓
Read PDFs—✓
Hear audio——
Watch video——
Call tools✓✓
Force a tool call?Partial
Return strict JSONPartial✓
Search the web??
Run code??
Use a computer??

What breaks when you switch

Devstral Small (Jul '25) → Mistral Medium 3.5
Memory (4-bit): 14 GB → 77 GB
Mistral Medium 3.5 → Devstral Small (Jul '25)
Context: 262K → 256K
No image input
No PDF input