Magistral Medium 2506 vs Mistral Medium 3.5
Different efforts: medium · high
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Magistral Medium 2506—
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
Magistral Medium 2506—
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
Magistral Medium 2506—
Mistral Medium 3.56.3
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Magistral Medium 25061,280
Mistral Medium 3.51,400
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Magistral Medium 2506—
Mistral Medium 3.564.9
| Benchmark | Magistral Medium 2506 | Mistral Medium 3.5 |
|---|---|---|
| Humanity's Last Exam | — | 13.8 |
| Terminal-Bench 4.0 | — | 0 |
| AutomationBench-AA | — | 6.3 |
| Arena Text · creative writing | 1,280 | 1,400 |
| MMMU-Pro (vision) | — | 64.9 |
Specs
| Spec | Magistral Medium 2506 | Mistral Medium 3.5 |
|---|---|---|
| Input, per 1M | — | $1.50 |
| Output, per 1M | — | $7.50 |
| Cost per task? | — | $0.50 |
| Speed, tok/s | — | 155 |
| Context | — | 262K |
| Weights | Not verified | Open |
| Released | — | Apr 2026 |
| Memory (4-bit) | — | 77 GB |
| Licence | — | custom licence |
| Commercial use | — | Restricted |
| See images | ? | ✓ |
| Read PDFs | ? | ✓ |
| Hear audio | ? | — |
| Watch video | ? | — |
| Call tools | ? | ✓ |
| Force a tool call | ? | Partial |
| Return strict JSON | ? | ✓ |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Magistral Medium 2506 → Mistral Medium 3.5
Nothing known.
Mistral Medium 3.5 → Magistral Medium 2506