Mistral Medium 3 — reviews, specs & pricing
Mid-size multimodal Mistral model priced for high-volume enterprise deployment.
Summary
An enterprise-focused multimodal model that delivers frontier-level coding, reasoning and vision quality at a fraction of flagship pricing. Mistral positions it as roughly 90% of top-tier performance for far less money.
Sample use case
Cost-efficient enterprise reasoning, coding and vision tasks.
Specifications
- Provider: mistral
- License: closed
- Context: 128k tokens
- Input price: $0.4/M tok
- Output price: $2/M tok
- Released: 2025-05-07
Average rating 3.0 from 5 community reviews on Reviuws.
Community reviews
VentureBeat on Mistral Medium 3
Rating: 3.0 / 5 — by Carl Franzen (use case: enterprise assistant deployment)
VentureBeat frames Medium 3 as Mistral's aggressive push into enterprise AI alongside Le Chat Enterprise, a cost-competitive alternative despite Mistral's weaker enterprise mindshare.
Pros: Aimed squarely at enterprise use with the Le Chat Enterprise bundle; competitive cost.
Cons: Still trails leaders in enterprise and developer mindshare.
TechCrunch on Mistral Medium 3
Rating: 3.0 / 5 — by Kyle Wiggers (use case: cost-efficient API access)
TechCrunch reports Mistral's claim that Medium 3 delivers leading performance for its price, emphasising API pricing of $0.40 and $2 per million input and output tokens.
Pros: Strong claimed performance per dollar; low published API pricing.
Cons: Performance claims are self-reported and not independently verified here.
SiliconANGLE on Mistral Medium 3
Rating: 3.0 / 5 — by Kyt Dotson (use case: general enterprise LLM use)
SiliconANGLE covers the Medium 3 release within Mistral's model family, reiterating the low-cost, high-performance positioning against larger proprietary competitors without hands-on testing.
Pros: Low reported cost for its performance tier; coherent model family.
Cons: Launch-announcement coverage rather than a tested verdict.
the-decoder on Mistral Medium 3
Rating: 3.0 / 5 — by Matthias Bastian (use case: European enterprise AI)
Bastian frames Medium 3 and Le Chat Enterprise as Mistral strengthening its position as Europe's leading AI startup, describing a medium-sized model competitive with much larger ones.
Pros: Competitive against larger models despite its size; paired with a new enterprise assistant.
Cons: Performance-versus-size claims not independently benchmarked.
16x Eval on Mistral Medium 3
Rating: 3.0 / 5 — by Zhu Liang (use case: coding and writing evaluation)
An independent hands-on evaluation compares Medium 3 with GPT-4.1, Claude 3.7 Sonnet, DeepSeek V3 and Gemini 2.5 Pro on a coding task and writing tasks, giving concrete comparative scores rather than relying on vendor claims.
Pros: Handles straightforward coding competently against strong peers; reproducible method.
Cons: Not a top performer against leaders like Claude 3.7 Sonnet on every task.
Frequently asked questions
What is Mistral Medium 3?
An enterprise-focused multimodal model that delivers frontier-level coding, reasoning and vision quality at a fraction of flagship pricing. Mistral positions it as roughly 90% of top-tier performance for far less money.
How much does Mistral Medium 3 cost?
Mistral Medium 3 costs $0.4 per million input tokens and $2 per million output tokens.
Is Mistral Medium 3 open source?
Mistral Medium 3 is released under the closed license.