Skip to main content
AI Choice Engine

Model comparison

Mistral Large 3 vs Mistral Small 4

Two Mistral tiers compared on the figures that decide which one a workload actually needs.

Catalog record checked September 26, 2026Individual provider fields may changeEvidence confidence: High — see receipts below

Mistral

Mistral Large 3

Frontier · Open weights

vs

Mistral

Mistral Small 4

Budget · Open weights

AI model capability comparison
SpecificationMistral Large 3Mistral Small 4
ProviderMistralMistral
TierFrontierBudget
Context window262K262K
Max outputNot verifiedUnverifiedNot verifiedUnverified
Input / 1M tokens$0.50Winner: $0.15
Output / 1M tokens$1.50Winner: $0.60
WeightsOpenOpen
Parameters675B total / 41B active (sparse MoE)119B total / 6.5B active (MoE)
Reasoning levelsNot verifiedUnverifiedNot verifiedUnverified
Modalitiestext, imagetext, image
LicenseApache 2.0Apache 2.0
API model idmistral-large-2512mistral-small-2603
ReleasedDecember 2, 2025March 16, 2026
Artificial Analysis Intelligence Index (2026-09-26)9.3Winner: 11.3
GPQA Diamond (2026-09-26)68Not verifiedUnverified

Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Bold indicates the better figure where one is objectively better. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.

Benchmark receipts

Where each score comes from, and how far it can be compared across models.

  • 2026-09-26
    Artificial Analysis GPQA Diamond (independent; replaces Mistral's non-comparable base-model 5-shot 43.9)

    Graduate-level science reasoning: multiple-choice questions written by domain PhDs.

    Comparable with caveatSmall question pool: differences under ~2 points are within run-to-run noise. Chain-of-thought vs direct answering must match to compare. Saturated: frontier models score ~93–96% and Artificial Analysis dropped it from its Intelligence Index in v4.3 — use Humanity's Last Exam for hard reasoning.

  • 2026-09-26
    Artificial Analysis (Reasoning)

    Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.

    Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check measuredAt.

Pricing tiers

Mistral Large 3: Mistral first-party API $0.50/$1.50 per MTok; cached input $0.05/MTok. Open weights (Apache 2.0), EU-hosted.

Mistral Small 4: Mistral first-party API $0.15/$0.60 per MTok. Open weights (Apache 2.0). 119B total / 6.5B active MoE.

FrontierOpen weightsRecord checked September 26, 2026

Mistral Large 3

Mistral Large 3 is Mistral's December 2025 open-weight (Apache 2.0) 675B/41B sparse MoE with a 262K context at $0.50/$1.50 — Mistral now features Medium 3.5 as its flagship.

Best for

  • Open-weight deployments
  • Multimodal EU-hosted inference
  • Balanced frontier work

Watch out

No longer Mistral's featured flagship (Medium 3.5 is) and far behind it on the Artificial Analysis index; max output is not published. Verify multimodal support and EU data residency on your endpoint.

BudgetOpen weightsRecord checked September 26, 2026

Mistral Small 4

Mistral Small 4 is Mistral's low-cost Apache-2.0 MoE — 6.5B active parameters with a 256K context and image input.

Best for

  • Cheap EU-hosted inference
  • Open-weight self-hosting
  • High-volume extraction

Watch out

Small active size: not for hard reasoning or agentic coding; max output is not published.

When the cheaper one wins

Mistral Small 4 is cheaper on output at $0.60 per million tokens against $1.50 for Mistral Large 3 — about 2.5×. Use the cheaper tier for classification, extraction, summarisation, and any task where the expensive model’s extra score does not change the accepted output. The expensive one only pays if your hardest task actually fails on the cheap tier. These are standard-tier API rates, excluding batch and cache discounts.

Run the model picker

Evidence confidence: High

How strong and complete the evidence behind this comparison is — not a prediction of which model is better.

  • Pricing verified on both sides — Input and output rates are verified for both models.
  • 3/5 core specs verified on both sides — Not published for at least one side: max output, reasoning levels.
  • 1 shared named benchmark with differing scores — Measured on: Artificial Analysis Intelligence Index.
  • Verified within the last 90 days — Newest catalog check was 2 days ago.
  • Both models carry source citations — Each side has at least two catalog sources on record.

Source receipts

Each catalog figure was checked against the provider or an independent second source on the date shown.

Related comparisons

Diving deeper on one model? Mistral Large 3 · Mistral Small 4

Common questions

Mistral Large 3 vs Mistral Small 4

Answered from the verified figures on this page rather than general guidance.

Is Mistral Large 3 or Mistral Small 4 cheaper for input?
Mistral Small 4 is cheaper at $0.15 per million input tokens, against $0.50 for Mistral Large 3 — roughly 3.3× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; Mistral Large 3 has tiered pricing: Mistral first-party API $0.50/$1.50 per MTok; cached input $0.05/MTok. Open weights (Apache 2.0), EU-hosted. Mistral Small 4 has tiered pricing: Mistral first-party API $0.15/$0.60 per MTok. Open weights (Apache 2.0). 119B total / 6.5B active MoE.
Is Mistral Large 3 or Mistral Small 4 cheaper for output?
Mistral Small 4 is cheaper at $0.60 per million output tokens, against $1.50 for Mistral Large 3 — roughly 2.5× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; Mistral Large 3 has tiered pricing: Mistral first-party API $0.50/$1.50 per MTok; cached input $0.05/MTok. Open weights (Apache 2.0), EU-hosted. Mistral Small 4 has tiered pricing: Mistral first-party API $0.15/$0.60 per MTok. Open weights (Apache 2.0). 119B total / 6.5B active MoE.
Which has the larger context window, Mistral Large 3 or Mistral Small 4?
Both accept about 262K tokens of context, so document length will not decide between them.
Should I use Mistral Large 3 or Mistral Small 4?
Mistral Large 3 is the frontier tier and Mistral Small 4 the budget tier. The useful question is whether your hardest task actually fails on the cheaper one — most production volume such as classification, extraction and summarisation does not.

Next step

Choosing between them

The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.

  • Input price: Mistral Small 4 costs $0.15 per 1M tokens versus $0.50 for Mistral Large 3 — a 3.3x difference at the headline tier.
  • Positioning: Mistral Large 3 sits in the frontier tier, Mistral Small 4 in the budget tier — most production volume (classification, extraction, summarisation) does not need the pricier tier.
Run the model pickerEstimate stack costAll model comparisons

Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.