The AI arena is free today

Open Superagent

Ministral 3 (8B Instruct 2512) vs Mistral Large 3

Mistral Large 3 significantly outperforms across most benchmarks.

Mistral AI · Mistral AI · Updated for 2026

Which is better?

Ministral 3 (8B Instruct 2512) outperforms in 0 benchmarks, while Mistral Large 3 is better at 4 benchmarks (Arena Hard, MATH, MM-MT-Bench, Wild Bench). Mistral Large 3 significantly outperforms across most benchmarks.

Based on current benchmark, pricing, and model metadata for 2026.

Choose Ministral 3 (8B Instruct 2512)

  • you want the most recent training data — it shipped Dec 2025

Choose Mistral Large 3

  • you want the strongest raw capability — it leads on 4 of 4 shared benchmarks

At a glance

The differences that matter most.

Benchmark wins
0 of 4
4 of 4
Input price
— / M
$2.00 / M
Output price
— / M
$5.00 / M
Context window
128,000
Released
Dec 2025
Sep 2025
License
Apache 2.0
Apache 2.0

Performance Benchmarks

Comparative analysis across standard metrics

4 benchmarks

Ministral 3 (8B Instruct 2512) outperforms in 0 benchmarks, while Mistral Large 3 is better at 4 benchmarks (Arena Hard, MATH, MM-MT-Bench, Wild Bench).

Mistral Large 3 significantly outperforms across most benchmarks.

Tue Aug 25 2026 • llm-stats.com

Arena Performance

Playground indexes and blind preference scores

Model Size

Parameter count comparison

667.0B diff

Mistral Large 3 has 667.0B more parameters than Ministral 3 (8B Instruct 2512), making it 8337.5% larger.

Mistral AI
Ministral 3 (8B Instruct 2512)
8.0Bparameters
Mistral AI
Mistral Large 3
675.0Bparameters
8.0B
Ministral 3 (8B Instruct 2512)
675.0B
Mistral Large 3

Context Window

Maximum input and output token capacity

Only Mistral Large 3 specifies input context (128,000 tokens). Only Mistral Large 3 specifies output context (8,192 tokens).

Mistral AI
Ministral 3 (8B Instruct 2512)
Input- tokens
Output- tokens
Mistral AI
Mistral Large 3
Input128,000 tokens
Output8,192 tokens
Tue Aug 25 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both Ministral 3 (8B Instruct 2512) and Mistral Large 3 support multimodal inputs.

They are both capable of processing various types of data, offering versatility in application.

Ministral 3 (8B Instruct 2512)

Text
Images
Audio
Video

Mistral Large 3

Text
Images
Audio
Video

License

Usage and distribution terms

Both models are licensed under Apache 2.0.

Both models share the same licensing terms, providing consistent usage rights.

Ministral 3 (8B Instruct 2512)

Apache 2.0

Open weights

Mistral Large 3

Apache 2.0

Open weights

Release Timeline

When each model was launched

Ministral 3 (8B Instruct 2512) was released on 2025-12-04, while Mistral Large 3 was released on 2025-09-01.

Ministral 3 (8B Instruct 2512) is 3 months newer than Mistral Large 3.

Ministral 3 (8B Instruct 2512)

Dec 4, 2025

8 months ago

3mo newer
Mistral Large 3

Sep 1, 2025

11 months ago

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against Ministral 3 (8B Instruct 2512) and Mistral Large 3 side-by-side, then vote on the output you prefer.

Ministral 3 (8B Instruct 2512)
✓ Preferred
Mistral Large 3
Open in Playground

FAQ

Common questions about Ministral 3 (8B Instruct 2512) vs Mistral Large 3.

Which is better, Ministral 3 (8B Instruct 2512) or Mistral Large 3?

Mistral Large 3 significantly outperforms across most benchmarks. Ministral 3 (8B Instruct 2512) is made by Mistral AI and Mistral Large 3 is made by Mistral AI. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does Ministral 3 (8B Instruct 2512) compare to Mistral Large 3 in benchmarks?

Ministral 3 (8B Instruct 2512) scores MATH: 87.6%, Wild Bench: 66.8%, Arena Hard: 50.9%, MM-MT-Bench: 8.1%. Mistral Large 3 scores MATH: 90.4%, MM-MT-Bench: 84.9%, MMLU-Redux: 82.0%, TriviaQA: 74.9%, MMMLU: 74.2%.

What are the context window sizes for Ministral 3 (8B Instruct 2512) and Mistral Large 3?

Ministral 3 (8B Instruct 2512) supports an unknown number of tokens and Mistral Large 3 supports 128K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.