Skip to main content
Watch: I Built an AI App Without Leaving Databricks

Model comparison - Same provider, successor

Mistral Large 3 vs Mistral Large 4

Mistral Large 3 and Mistral Large 4 are both Mistral AI models. On list price, Mistral Large 4 charges 2.7x Mistral Large 3's input price ($1.36 vs $0.50 per 1M tokens), and Mistral Large 4 charges 2.8x Mistral Large 3's output price ($4.18 vs $1.50 per 1M tokens). Mistral Large 3 has a 256,000-token context window and Mistral Large 4 1,000,000 tokens. Mistral Large 3 was released on December 1, 2025 and Mistral Large 4 on October 5, 2026. Mistral Large 3 publishes open weights (Apache 2.0); Mistral Large 4 is closed-weights.

Last verified from each provider's pricing page. All sources

Mistral AI - Open

Mistral Large 3
Input / 1M
$0.50
Output / 1M
$1.50
Context
256K

Mistral AI - Closed

Mistral Large 4
Input / 1M
$1.36
Output / 1M
$4.18
Context
1M

Side-by-side specs

SpecMistral Large 3Mistral Large 4
ProviderMistral AIMistral AI
Input / 1M tokens$0.50$1.36
Output / 1M tokens$1.50$4.18
Cache read / 1M$0.05$0.14
Batch input / 1M$0.25$0.68
Batch output / 1M$0.75$2.09
Context window256,0001,000,000
Release date
WeightsOpenClosed
LicenseApache 2.0-
API model idmistral-large-2512mistral-large-4
AI Gateway idmistral/mistral-large-3mistral/mistral-large-4
Last verified

Prices in USD per 1M tokens at each provider's standard first-party rate. A dash means the provider does not bill that item separately or no primary page states it.

What each costs in practice

The same three workloads every model page uses, worked from each model's standard list rates. Swap in your own token counts with the math on each model's page.

WorkloadMistral Large 3Mistral Large 4
One agent session1M input tokens and 100K output tokens, no caching.$0.65$1.78
A chat app10,000 requests a day at 2K input and 500 output tokens each.$17.50$525.00 per 30-day month$48.10$1,443.00 per 30-day month
The same session at a 90% cache hit90% of the input is read from the prompt cache. The one-time cache write premium is not included.$0.245$0.68
Show the math

Mistral Large 3

  • One agent sessionInput: 1M x $0.50 = $0.50Output: 0.1M x $1.50 = $0.15
  • A chat appInput: 20M x $0.50 = $10.00Output: 5M x $1.50 = $7.50
  • The same session at a 90% cache hitCached input: 0.9M x $0.05 = $0.045Uncached input: 0.1M x $0.50 = $0.05Output: 0.1M x $1.50 = $0.15

Mistral Large 4

  • One agent sessionInput: 1M x $1.36 = $1.36Output: 0.1M x $4.18 = $0.418
  • A chat appInput: 20M x $1.36 = $27.20Output: 5M x $4.18 = $20.90
  • The same session at a 90% cache hitCached input: 0.9M x $0.14 = $0.126Uncached input: 0.1M x $1.36 = $0.136Output: 0.1M x $4.18 = $0.418

Benchmarks under the same setting

No board in the benchmark directory lists both Mistral Large 3 and Mistral Large 4 under the same setting, so there is no like-for-like score to show here.

Set them up in your tools

Step-by-step guides whose official docs list Mistral Large 3 or Mistral Large 4 by model id.

Model pages

FAQ

How do Mistral Large 3 and Mistral Large 4 compare on price?

On list price, Mistral Large 4 charges 2.7x Mistral Large 3's input price ($1.36 vs $0.50 per 1M tokens), and Mistral Large 4 charges 2.8x Mistral Large 3's output price ($4.18 vs $1.50 per 1M tokens). For one agent session of 1M input and 100K output tokens with no caching, that is $0.65 on Mistral Large 3 and $1.78 on Mistral Large 4. Prices are each provider's first-party list rates, verified October 17, 2026 and October 17, 2026.

What are the context windows of Mistral Large 3 and Mistral Large 4?

Mistral Large 3 has a 256,000-token context window and Mistral Large 4 1,000,000 tokens.

Are Mistral Large 3 and Mistral Large 4 open weights?

Mistral Large 3's weights are published at huggingface.co under the Apache 2.0. Mistral Large 4 is closed-weights and is called through Mistral AI's API as mistral-large-4.

More comparisons

Sources

Mistral Large 3 verified from docs.mistral.ai; Mistral Large 4 verified from docs.mistral.ai. This page is as current as the older of the two: .

Newsletter

Get model price changes as they ship

One email per week. New model releases, price cuts and what they mean for your bill.