Skip to content

Mistral modified_mit_license Released Apr 29, 2026 Updated

Mistral Medium 3.5

ToolsGuide Scores, rankings, test results and price.

Model summary

  • Text #140 / 416

    318 / 1,000

    ToolsGuide Score· votes 351, tests 266

  • Vision #89 / 167

    325 / 1,000

    ToolsGuide Score· votes only

  • Coding #130 / 147

    199 / 1,000

    ToolsGuide Score· votes only

  • Agents #52 / 53

    340 / 1,000

    ToolsGuide Score· votes 385, tests 291

  • Price per 1M tokens

    In $1.50· Out $7.50

    $1.50

    Input · median text model $0.43

Summary

Mistral Medium 3.5 is Mistral's model, released Apr 29, 2026. It's ranked in 4 categories and does best at text: #140 of 416, with a ToolsGuide Score of 318 against a median of 235.

It also ranks #89 of 167 for vision (325), #130 of 147 for coding (199), #52 of 53 for agents (340).

Its vision and coding score uses votes only, because it has published fewer than two of the tests used for those categories.

It costs $1.50 per million input tokens and $7.50 per million output tokens, about 3.4× the median text model ($0.43).

It has published results on 11 tests and has the top score on 3 of them, including Tau3 Airline and Tau3 Retail.

Specifications

Creator
Mistral
Licence
modified_mit_license
Released
Apr 29, 2026
Context window
256K tokens
Input price
$1.50 / 1M
Output price
$7.50 / 1M
Tests reported
11
Settings ranked
mistral-medium-3.5, Mistral Medium 3.5

Model comparison

Against the ten best text models.

Text score

ToolsGuide Score out of 1,000 · Higher is better

Vision score

ToolsGuide Score out of 1,000 · Higher is better

Input price

USD per 1M input tokens · Lower is better

Leaderboards

Where Mistral Medium 3.5 sits among the top 20 in each category.

Text

#140 of 416 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full text leaderboard →

Vision

#89 of 167 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full vision leaderboard →

Coding

#130 of 147 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full coding leaderboard →

Agents

#52 of 53 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full agents leaderboard →

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Score against price

Text · the 10 priced models ranked closest · up and to the left is better value

Against its nearest rivals

Text prompts by kind · ToolsGuide Score from votes · hover for values

Text

mistral-medium-3.5 · hover a bar for its rank

Vision

Mistral Medium 3.5 · hover a bar for its rank

Coding

Mistral Medium 3.5 · hover a bar for its rank

Test results

Every test Mistral Medium 3.5 has a published result on, out of 100%, with its rank among the models that took it.

  • AIME 2025

    #64 / 123

    Competition maths problems from the 2025 American Invitational Mathematics Exam.

    86.3%

    Self-reported

  • SWE-Bench Verified

    #27 / 120

    Fixing real bugs in real GitHub projects, checked by the projects' own tests.

    77.6%

    Self-reported

  • BrowseComp

    #55 / 68

    Finding hard-to-locate facts by browsing the web.

    48.6%

    Self-reported · In the agents score

  • IFBench

    #33 / 44

    Instruction Following Benchmark evaluating model's ability to follow complex instructions

    69.0%

    Self-reported

  • Finance Agent v2

    #24 / 28

    Finance Agent v2 is an agentic financial-analysis benchmark from Vals that evaluates models on real-world finance workflows, measuring their ability to retrieve and reason over financial documents, perform multi-step calculations, and produce accurate analyses.

    32.1%

    Self-reported

  • Tau3 Banking

    #10 / 13

    τ³-Bench banking domain evaluates agentic models on multi-turn, tool-using customer-support scenarios in a simulated retail banking environment.

    13.4%

    Self-reported

  • COLLIE

    #4 / 11

    COLLIE is a grammar-based framework for systematic construction of constrained text generation tasks. It allows specification of rich, compositional constraints across diverse generation levels and modeling challenges including language understanding, logical reasoning, and semantic planning. The COLLIE-v1 dataset contains 2,080 instances across 13 constraint structures.

    95.8%

    Self-reported

  • Beyond AIME

    #6 / 7

    Beyond AIME is a difficult mathematical reasoning benchmark designed to test deeper reasoning chains and harder decomposition than standard AIME-style problem sets.

    66.9%

    Self-reported

  • Tau3 Airline

    #1 / 1

    τ³-Bench airline domain evaluates agentic models on multi-turn, tool-using customer-support scenarios in a simulated airline booking and reservations environment.

    72.0%

    Unverified

  • Tau3 Retail

    #1 / 1

    τ³-Bench retail domain evaluates agentic models on multi-turn, tool-using customer-support scenarios in a simulated online retail environment.

    76.1%

    Unverified

  • Tau3 Telecom

    #1 / 1

    τ³-Bench telecom domain evaluates agentic models on multi-turn, tool-using customer-support and troubleshooting scenarios in a simulated telecommunications environment.

    91.4%

    Unverified