Mistral

Ministral 3 14B

Dec 2, 2025Open weightsModels.devHugging Face

Compare

Intelligence

6.0

#630 of 744Rank 630 of 744 measured models

Coding

14.4

#249 of 298Rank 249 of 298 measured models

Output speed

64.8 t/s

#264 of 403Rank 264 of 403 measured models

Price/1M

$0.20

#161 of 659Rank 161 of 659 measured models

Context window

256K

#518 of 1199Rank 518 of 1199 measured models

Cost per task

$0.0194

#22 of 193Rank 22 of 193 measured models

Ministral 3 14B scores 6.0 on the Artificial Analysis Intelligence Index, ranking 630th of 744 measured models. The closest models on this index are DeepSeek-Coder-V2 (6.0), Granite 4.0 H Small (6.0), and Hermes 3 - Llama-3.1 70B (6.0). At $0.20 per million tokens (3:1 input/output blended price), it is cheaper than Hermes 3 - Llama-3.1 70B ($0.70), more expensive than Granite 4.0 H Small ($0.107). With an output speed of 64.8 t/s, it is faster than Granite 4.0 H Small (14.4 t/s) and Hermes 3 - Llama-3.1 70B (31.5 t/s).