Advertise here
#240

Mistral Large 4

Mistral Release: 2026-10-06 Tested on: 2026-10-06 19:26 mistralai/mistral-large-4-0::medium
1.05T total (49B active)MoEClosed
(high) (medium) (low) (none)

Summary

Mistral Large 4 scores 6.2 on AI BENCHY and ranks #240. It has 9.2 reliability, a 55.1% pass rate, $1.456 total cost, and 265.03s average response time.

What makes Mistral Large 4 unique: It stands out most in Combined, where it ranks #3, while General Intelligence is its weakest area at #16. It uses unusually many reasoning tokens, which can help explain its slower or more expensive runs.

Model facts

Researched on 2026-10-06

Reported
Parameters
1.05T total (49B active)
Architecture
MoE
Availability
Closed
License
-

Mistral reports 1.05T total and 49B active parameters, plus a 1.6B vision encoder. The vendor describes this public preview as open-weight, but its checkpoint page currently supplies no weight download or license, and no Large 4 checkpoint was found in the official Hugging Face registry. Classified as API-only until the exact weights and license can be verified. The OpenRouter route exposes 524,288 context tokens and 262,144 completion tokens, with tools and structured outputs but no configurable reasoning parameter.

Score

6.2

Consistency

7.6

Total Output Tokens

627,233

Total Input Tokens

212,024

Input Price

$0.680 / 1M

Output Price

$2.090 / 1M

Cache Read Price

$0.070 / 1M

Cache Write Price

N/A

Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.

Tests Correct

Wrong Tests: 14

Attempt pass rate: 55.1%

Flaky tests

7

Flaky tests had mixed outcomes across runs (at least one pass and one fail).

Response Time (avg)

265.03s

Response Time (max): 1760.20s

Response Time (total): 6095.69s

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#240 Mistral Large 4

medium
Cost
$0.011
Time
63.4s
Tokens
5,291 tok

Charts

Choose the first model, then click a second model to open a side-by-side page.

Total Output Tokens

Score vs Total Output Tokens

Quick Compare

Category Breakdown

Category Score Consistency Tests Correct
Agentic 6.1 3.1
Anti-AI Tricks 8.3 10.0
Coding 3.4 7.2
Combined 7.3 5.8
Data parsing and extraction 10.0 10.0
Domain specific 2.9 7.2
General Intelligence 3.5 2.2
Instructions following 8.3 10.0
Puzzle Solving 6.1 4.7
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models