Which model should you pick?
Filter by legal jurisdiction, budget, context and vision. The ranking uses published prices, each model official positioning, context size and the speed our probes actually measure.
Ranking
| # | Version | Provider | Data jurisdiction | Price $/M (in / out) | Measured speed (first token) | Context | Our probe | Score |
|---|---|---|---|---|---|---|---|---|
| 1 | Mistral Small 4 | Mistral AI | EU | 0,15 / 0,6 USD | 244 ms | 262 k | Operational | 88,3 |
| 2 | Mistral Large 4 | Mistral AI | EU | 0,7 / 2,1 USD | 241 ms | 1 M | Operational | 57,1 |
| 3 | Mistral Medium 3.5 | Mistral AI | EU | — | 246 ms | — | Operational | 26,8 |
| 4 | Kolibri Kolibri A3B | Aleph Alpha | EU | — | — | — | Not probed (no key) | 23,5 |
| 5 | LightOnOCR LightOnOCR-2 | LightOn | EU | — | — | — | Not probed (no key) | 20,5 |
Why this ranking
1. Mistral Small 4 — Mistral AI (EU) · Instruct + raisonnement + code en un modèle
Weighting : Cost 100 · Speed 99 · Capability 62 · Context 60 · Our probe 100
2. Mistral Large 4 — Mistral AI (EU) · Multimodal, agents, QnA documentaire
Weighting : Cost 28 · Speed 100 · Capability 85 · Context 99 · Our probe 100
3. Mistral Medium 3.5 — Mistral AI (EU) · Déprécié
Weighting : Cost 0 · Speed 98 · Capability 50 · Context 20 · Our probe 100
Score 0-100: cost (input price 30%, output 70%), capability (official positioning in the range), context (log scale), speed (measured first token; neutral 45/100 when not probed), state (100 operational, 55 degraded, 0 down, 60 not measured).