Mistral Large 2 / Mixtral family

https://mistral.ai/news/mistral-large-2/

French frontier-model lab (Paris). Family includes Mistral Large 2 (123B dense, Jul-2024), Mixtral 8x22B (open-weights MoE), Mistral Small 3 (24B, open-weights Jan-2025), Codestral (code specialist). EU-headquartered alternative to US/China model labs; GDPR-friendly data residency. Distinct from the Mistral Le Chat memory feature (already in catalog) — this row covers the model family itself as substrate.

At a glance

Type
Frontier model family — Mistral Large 2 + Mixtral 8x22B + Mistral Small 3
Tier
T1
Created
2023-04 (Mistral AI founded); 2023-09 (Mistral 7B); 2023-12 (Mixtral 8x7B); 2024-04 (Mixtral 8x22B); 2024-07 (Large 2); 2025-01 (Small 3)
Latest release
Mistral Large 2.1 (2025-Q1); Mistral Small 3 (2025-01); Codestral 2 (2025)
License
Apache 2.0 (Mixtral / Small 3 / Codestral); proprietary (Large 2 / Pixtral Large)
Pricing
API: Mistral Large 2 $2/$6 per 1M; Small 3 free / $0.10-$0.30; Apache 2.0 open weights for Mixtral + Small + Codestral; Enterprise Mistral Forge on-prem
Funding
€2.4B+ total raised (~$2.7B USD); €11.7B valuation Series C Sept-2025 (ASML lead); a16z, General Catalyst, DST, Lightspeed, Nvidia investors

Taxonomy

storage
parametric
retrieval
parametric-recall
persistence
parametric-permanent
update
read-only
unit
weight
governance
opaque
conflict
n/a

When to use

Optimised for: EU-hosted enterprise; sovereign AI for European governments; cost-effective hybrid (open + commercial) family

Anti-fit: not for users needing absolute frontier capability (lags Claude/GPT/Gemini on most benchmarks); not for non-Apache OSS users who want weight openness on Large 2

Pros & cons

Pros

Strongest EU AI-sovereignty story (Paris-HQ, GDPR-native, French gov customer); Apache 2.0 open weights for Mixtral / Small / Codestral; €2.7B raised at €11.7B valuation.

Cons

Trails US/China frontier labs on raw benchmarks; Large 2 not open-weights; smaller training-compute budget; multimodal less mature than peers.

Claims & capabilities

Mistral Large 2: 123B dense, 128k ctx, MMLU 84%; Mixtral 8x22B open-weights (Apache 2.0); Mistral Small 3 24B open-weights matches Llama 3.3 70B; EU-hosted by default; €11.7B valuation Sept-2025

Technical surface

API surface
REST + SDK (Python, JS); HuggingFace Transformers for open-weights; Azure AI Foundry; AWS Bedrock; GCP Vertex
Backend storage
not applicable — substrate foundation model
Deployment
Managed cloud (La Plateforme) + on-prem (Mistral Forge enterprise) + open-weights (Apache 2.0 for Mixtral + Small + Codestral) + Azure / AWS / GCP hosted
Embedding model
not applicable — not a memory product
Multi-tenancy
not applicable — substrate foundation model
MCP
not applicable — substrate foundation model
A2A
not applicable — substrate foundation model
OpenTelemetry
not applicable — substrate foundation model

Compare Mistral Large 2 / Mixtral family with…

Similar systems

Other foundation models (substrate reference) in the catalog, ranked by inbound references.

  • Anthropic Claude (foundation models) T1

    Claude family of frontier foundation models — Sonnet 4 (workhorse), Opus 4.5 (frontier capability), Haiku 4.5 (fast/cheap). Frontier model used as substrate by Claude Code, Anthropic Memory tool, Cline, Aider, Continue.dev, Cursor, Goose, and dozens of other catalog entries. Memory in Claude is parametric (in weights) — discrete episodic memory lives outside the model in the Claude Memory tool (see Platform-provider memory section).

  • OpenAI GPT family (GPT-5 / GPT-4o / o3 / o4) T1

    OpenAI's foundation model family — GPT-5 (2025-08 flagship), GPT-4o (multimodal workhorse), o3/o4 reasoning series. Substrate for ChatGPT, the OpenAI API, Microsoft Copilot, Azure OpenAI, and many downstream agents in this catalog (Devin, Codex CLI, Operator, Cursor mode). Parametric memory in weights; discrete user memory ships in ChatGPT Memory feature (separate row).

  • Google Gemini 3 family T1

    Google DeepMind's Gemini 3 frontier model family — Gemini 3 Pro (top tier), 3 Flash (fast), 3 Nano (on-device). Substrate for Google Workspace AI, Vertex AI Agent Platform, Gemini Code Assist, Project Mariner, and Project Astra. Multimodal-native; 1M-10M context window. Largest context window of any frontier model. Used by Gemini Enterprise Agent Platform Memory Bank (separate row).

  • Meta Llama 4 family T1

    Meta's Llama 4 family (released 2025-04) — Scout (109B MoE, 17B active), Maverick (400B MoE, 17B active), Behemoth (2T MoE, 288B active; preview). First Llama generation to use mixture-of-experts and to be multimodal-native. Substrate for many OSS agents (Open-Interpreter, Ollama-served local agents, vLLM-served enterprise inference). Llama community license (commercial use permitted under 700M MAU threshold).

  • xAI Grok 4 T1

    Elon Musk's xAI frontier model — Grok 4 launched 2025-07. Trained on Colossus (100k+ GPU supercluster in Memphis). Deeply integrated with X/Twitter (real-time tweet data access). Grok 4 Heavy variant for premium tier. No frontier-tier open-weights — Grok 1 (314B MoE) is the only weight release.

  • 01.AI Yi family T2

    Kai-Fu Lee's 01.AI (founded 2023, Beijing). Yi family open-weights — Yi-34B (2023), Yi-Large (proprietary 2024), Yi-Lightning (2024-10, frontier-tier, matched GPT-4o on LMSYS chatbot arena). Apache 2.0 (Yi-34B and base sizes). Important second Chinese open-weights option alongside DeepSeek + Qwen.

Related systems

References (2)

  • Azure Machine Learning depends on at runtime — adjacent-infrastructure cell: Mistral Le Chat consumer (separate row); Codestral for code agents; Mistral Forge on-prem; AWS Bedrock / Azure AI distribution
  • Bedrock AgentCore (AWS) depends on at runtime — adjacent-infrastructure cell: Mistral Le Chat consumer (separate row); Codestral for code agents; Mistral Forge on-prem; AWS Bedrock / Azure AI distribution

Referenced by (3)

  • Mistral Agents API depends on at runtime — adjacent-infrastructure cell: self-contained Mistral platform
  • Mistral Embed depends on at runtime — al-embed; available via Mistral La Plateforme and Azure AI.
  • Mistral Vibe (Remote Agents) + Mistral Medium 3.5 depends on at runtime — adjacent-infrastructure cell: requires Mistral la Plateforme + Le Chat; integrates GitHub/Linear/Jira/Sentry/Slack/Teams; NVIDIA NIM containers for self-host

Row last verified 2026-05-14. Catalog data is CC-BY-4.0 — see how to read this.