[good]

Mistral

Mistral Large 3

4.6
out of 10

Mistral's December 2025 flagship and the most commercially permissive large model in this comparison. Mistral Large 3 is released under Apache 2.0 — genuinely open for commercial use without royalties or usage restrictions. At 675B total parameters with 41B active per token (mixture-of-experts), it scores 23 on the Artificial Analysis Intelligence Index at $0.50/$1.50 per 1M tokens. For enterprise teams that need open-weight licensing terms, the math is straightforward: comparable capability to other open-weight models, completely unrestricted commercial use, and a 256K context window that covers most document workflows.

Context window

256K tokens

API (blended)

$0.75/1M

Consumer access

Free (limited) / $15/mo

Multimodal

Yes

Strengths

  • +Apache 2.0 license — fully open commercial use, no royalties, no usage restrictions
  • +Competitive pricing at $0.75/1M blended — budget tier API
  • +256K context window — handles most enterprise document workloads
  • +Mixture-of-experts architecture: only 41B parameters active per token, efficient at scale
  • +Multimodal: text + image input

Weaknesses

  • -AA Index 23 is noticeably below frontier models — real reasoning gap vs Claude, GPT-5.2, Gemini 3 Pro
  • -Le Chat consumer product is less polished than ChatGPT or Gemini
  • -Slower than comparable open-weight models at 56 t/s
  • -Full 675B self-hosted deployment requires enterprise-grade multi-GPU infrastructure

Best for

enterprise open-weight deploymentcommercial applications requiring permissive licensingcost-sensitive API workloadsteams already using Mistral tooling

Not ideal for

cutting-edge reasoning tasks (use frontier models instead)users who need a polished consumer UIreal-time high-speed generation

Pricing details

Subscription plans

Free (Le Chat)Mistral Large 3 access via web with daily rate limits(Daily message cap; limited history retention)
Free
Le Chat ProHigher usage limits, longer history, priority access
$15/mo

API pricing

Mistral (la Plateforme)Standard tier via api.mistral.ai. Free trial credits available on signup. Batch API available.
$0.5/$1.5
OpenRouterSmall markup over direct Mistral pricing.
$0.52/$1.55
Self-hostedApache 2.0 — fully open commercial use. Download weights from HuggingFace. 675B model requires significant multi-GPU infrastructure; quantized versions available for smaller deployments.
Self-hosted

Prices verified February 2026. LLM pricing changes frequently — verify at the provider's site before budgeting.

Last updated: February 2026