DataLearner logo
MI

Mistral Large 3

Multimodal model

Mistral Large 3 (v25.12, Base/Instruct)

Release date: 2025-12-02Updated: 2026-07-17Views: 1,764
Live demoGitHubHugging FaceCompare
Parameters
675B
Context length
256K
Multilingual
No data
Reasoning ability
3/5

Mistral Large 3 is a multimodal model from MistralAI, released on 2025-12-02. It accepts text and image input and returns text output. The cataloged parameter count is 675B, with 41B active parameters per inference. The recorded context window is 256K. Cataloged capabilities include Reasoning model. The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Mistral Large 3

Model basics

Reasoning traces
Supported
Thinking modes
Standard ModeThinking Mode
Context length
256K tokens
Max output length
No data
Model type
Multimodal model
Modality (in / out)
Text, Image → Text
Release date
2025-12-02
Model file size
No data
MoE architecture
Yes
Total params / Active params
675B / 41B
Knowledge cutoff
No data
Mistral Large 3

Open source & experience

Code license
Apache 2.0
Weights license
Apache 2.0- Commercial use permitted
GitHub repo
N/A
Hugging Face
huggingface.co/collections/mistralai/mistral-la…
Live demo
console.mistral.ai
Mistral Large 3

Official resources

Paper
Paper
DataLearnerAI blog
N/A
Mistral Large 3

API details

API speed
3/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Learn about pricing modes
Standard
TypeConditionInputOutput
Text-$0.500/ 1M tokens$1.50/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

Mistral Large 3

Benchmark Results

Mistral Large 3 currently shows benchmark results led by Creative Writing (62 / 106, score 1409.30), Terminal Bench Hard (165 / 244, score 15.90), LiveCodeBench (177 / 251, score 46.50). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

1 evaluations
Benchmark / mode
Score
Rank/total
HLE
Standard Mode
4.20
515 / 565

General Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Standard Mode
68
338 / 463

Coding and Software Engineer

2 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
Standard Mode
46.50
177 / 251
SciCode
Standard Mode
36.60
113 / 131

Math and Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
AIME2025
Standard Mode
38
171 / 216

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
Standard Mode
1409.30
62 / 106

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
SimpleBench
Standard Mode
20.40
87 / 93

Agent Level Benchmark

3 evaluations
Benchmark / mode
Score
Rank/total
τ²-Bench - Telecom
Standard ModeTools
24.60
231 / 264
Terminal Bench Hard
Standard ModeTools
15.90
165 / 244
τ³-Banking
Standard ModeTools
5.80
150 / 167

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Standard Mode
36.20
241 / 282

Claw-style Agent Evaluation

3 evaluations
Benchmark / mode
Score
Rank/total
Claw Bench
Thinking ModeTools
78.60
22 / 29
Pinch Bench
Thinking ModeTools
72.20
29 / 38
PinchBench v2
unknown
54.48
37 / 45

AI Agent - Tool Usage

1 evaluations
Benchmark / mode
Score
Rank/total
Terminal-Bench 2.1
Standard ModeTools
12
174 / 196

Multimodal Understanding

2 evaluations
Benchmark / mode
Score
Rank/total
MMMU-Pro
Standard Mode
55.70
188 / 229
GDP.pdf
Standard Mode
1.20
113 / 119
View benchmark analysisCompare with other models
Mistral Large 3

Publisher

Mistral Large 3 (v25.12, Base/Instruct)

Model Overview

Mistral Large 3 is a multimodal model from MistralAI, released on 2025-12-02.

It accepts text and image input and produces text output. Its cataloged capabilities include Reasoning model. The cataloged parameter count is 675B, with 41B active parameters per inference. The recorded context window is 256K.

The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. The page records 2 API pricing rules from MistralAI; current provider pricing and conditions should be checked before deployment. The evaluation section contains 3 cataloged benchmark results with their recorded modes and scores. The page links 3 release, model-card, repository, or provider references for checking the underlying claims. Specifications, availability, and prices can change; undisclosed values are intentionally left unstated.

Mistral Large 3

FAQ

What is Mistral Large 3?

Mistral Large 3 is a multimodal model from MistralAI, released on 2025-12-02. It accepts text and image input and returns text output. The cataloged parameter count is 675B, with 41B active parameters per inference. The recorded context window is 256K. Cataloged capabilities include Reasoning model. The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Use the linked references to confirm current access, licensing, and provider-specific limits.

What input and output modalities does Mistral Large 3 support?

The current model record lists text and image as input and text as output.

What are the main recorded specifications for Mistral Large 3?

The cataloged parameter count is 675B, with 41B active parameters per inference. The recorded context window is 256K. Fields without a source-backed value remain undisclosed.

Does Mistral Large 3 have API pricing?

The page records 2 API pricing rules from MistralAI; current provider pricing and conditions should be checked before deployment.

Are benchmark results available for Mistral Large 3?

The evaluation section contains 3 cataloged benchmark results with their recorded modes and scores. Compare only results that use the same benchmark version and evaluation mode.

Is Mistral Large 3 open source?

The checkpoint is listed under the Apache 2.0 license, with the license link included in the model references. Review the linked license text before commercial or derivative use.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code