Meta Llama 4 family
https://ai.meta.com/blog/llama-4-multimodal-intelligence/
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).
Active · Verified 2026-10-31 · fresh as of 2026-10-31 · 36 cited facts, 1 with a third-party source · 1 recorded change · Suggest a correction
At a glance
- Tier
- T1
- Created
- 2023-02 (Llama 1); 2023-07 (Llama 2); 2024-04 (Llama 3); 2024-12 (Llama 3.3); 2025-04 (Llama 4) vendor fix
- Latest release
- No Llama release found after Llama 4 Scout + Maverick (2025-04-05); successor Muse family: Muse Spark (2026-04-08, closed), Muse Spark Open weights (2026-10-15) third-party fix
- Pricing
- Free for self-hosting (Llama community license, <700M MAU); hosted via Together $0.20-$3/M (Maverick); AWS Bedrock + Azure AI consumption vendor fix
- Freshness
- 2026-10-31 fix
- Layer
- large-models repo fix
- Licensing
- commercial vendor fix
When to use
Optimised for: open-weights frontier-tier deployers; cost-efficient inference via MoE; on-prem regulated deployments
Anti-fit: 700M MAU license trigger; not for users wanting weight transparency on dataset (Meta non-disclosing); Behemoth requires extreme GPU (2T MoE)
Pros & cons
Pros
Open weights (community license) — only frontier-tier family runnable on-prem; MoE architecture cheap to serve (17B active params); 10M context (Scout) matches Gemini 3.
Cons
License excludes 700M+ MAU products; dataset opacity; multimodal less mature than GPT-4o / Gemini 3; Behemoth not yet released as of 2025-12.
Claims & capabilities
Llama 4 Scout 109B MoE / 17B active / 10M context; Maverick 400B MoE / 17B active / 1M context; multimodal-native (text+image); Llama community license; first frontier open-weights to use MoE architecture
Technical surface
- API surface
- HTTP via Together / Bedrock / Azure; native via HuggingFace Transformers, vLLM, llama.cpp
- Backend storage
- not applicable — substrate foundation model
- Deployment
- Both (self-hosted + managed cloud) — open weights (llama.com / Hugging Face); hosted via AWS Bedrock, Azure AI, Together, Groq, Fireworks, Databricks
- 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
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).
- Mistral Large 2 / Mixtral family T1
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.
- 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. The x.ai site now brands itself SpaceXSI (page title, checked 2026-10-29).
- Cohere Command R+ / Command A T1
Toronto-based foundation-model lab (Aidan Gomez, ex-Google Brain 'Attention Is All You Need' co-author). Command family is enterprise / RAG-optimised: Command A (Mar-2025, 111B dense, 256k ctx), Command R+ (104B MoE), Command R7B (smallest tier). Announced merger with Aleph Alpha April-2026. Open weights via CC-BY-NC (non-commercial); Cohere-hosted for commercial.
Sources (35 cited facts, 1 third-party)
Each fact links to the page it came from; the badge says who published it (how to read sources).
- What it is Open-weights frontier model family (Llama 4 Scout / Maverick / Behemoth) vendor
- What it does & what's distinct Meta's Llama 4 family (released 2025-04) — Scout (109B MoE, 17B active), Maverick (400B … vendor
- Claims / benefits Llama 4 Scout 109B MoE / 17B active / 10M context; Maverick 400B MoE / 17B active / 1M c… vendor
- Modalities text, image (multimodal-native); audio via Llama-Omni research preview vendor
- Created 2023-02 (Llama 1); 2023-07 (Llama 2); 2024-04 (Llama 3); 2024-12 (Llama 3.3); 2025-04 (L… vendor
- Latest release No Llama release found after Llama 4 Scout + Maverick (2025-04-05); successor Muse famil… third-party
- License Llama 4 community license (Meta-custom; commercial OK <700M MAU) vendor
- GitHub github.com/meta-llama/llama (combined Llama repos); ~58k stars vendor
- Mindshare Dominant OSS frontier model — 1B+ HuggingFace downloads (Llama 3 family); de facto basel… vendor
- Funding Meta (META) public; ~$1.5T market cap; AI infra capex $60-65B 2025 vendor
- Customers / scale AWS Bedrock, Azure AI, Databricks, Together AI, Groq, Fireworks; downloads >1B (Llama 3 … vendor
- Pricing Free for self-hosting (Llama community license, <700M MAU); hosted via Together $0.20-$3… vendor
- Compliance License-only restrictions; deployers handle compliance (SOC 2 etc on hosting providers) vendor
- Data handling Open-weights — deployer-controlled vendor
- Deployment Both (self-hosted + managed cloud) — open weights (llama.com / Hugging Face); hosted via… vendor
- HQ Menlo Park, US (Meta FAIR + GenAI) vendor
- Founders / pedigree Meta Platforms (Mark Zuckerberg CEO); FAIR/GenAI led by Joelle Pineau / Ahmad Al-Dahle vendor
- Performance MMLU-Pro: Maverick ~80% (matches GPT-4o); LiveCodeBench: ~62%; GPQA Diamond: ~70%; Behem… vendor
- Code/weights release weights public on HuggingFace; community license vendor
- API surface HTTP via Together / Bedrock / Azure; native via HuggingFace Transformers, vLLM, llama.cpp vendor
- LLM lock open (Llama community license; many hosted providers) vendor
- Runtimes any (HuggingFace Transformers, vLLM, llama.cpp, MLX, Ollama) vendor
- Validated verticals OSS deployers, on-prem regulated industries (financial / health / gov), startups optimis… vendor
- Time to running <1hr (Ollama / vLLM / HuggingFace serve); <10min via hosted (Together / Bedrock) vendor
- Anti-fit 700M MAU license trigger; not for users wanting weight transparency on dataset (Meta non… vendor
- Optimised for open-weights frontier-tier deployers; cost-efficient inference via MoE; on-prem regulate… vendor
- Adjacent infrastructure HuggingFace Transformers; vLLM; llama.cpp; Together AI; Groq; AWS Bedrock; Azure AI Foun… vendor
- Pros Open weights (community license) — only frontier-tier family runnable on-prem; MoE archi… vendor
- Cons License excludes 700M+ MAU products; dataset opacity; multimodal less mature than GPT-4o… vendor
- Project & sources ai.meta.com/blog/llama-4-multimodal-intelligence/ vendor
- capability-benchmark-sources livecodebench: 43.4% (Llama 4 Maverick instruct, 2024-10-01 to 2025-02-01 window, 2025-0… repo
- architecture-deployment-locus hybrid vendor
- licensing-bucket commercial vendor
- positioning-platform-vs-product platform vendor
- taxonomy-three-layer-rollup large-models repo
What changed
Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)
- 2026-10-26
- Latest release “Llama 4 Scout + Maverick (2025-04); Behemoth in preview as of 2025” → “No Llama release found after Llama 4 Scout + Maverick (2025-04-05); successor Muse family: Muse Spark (2026-04-08, closed), Muse Spark Open weights (2026-10-15)” source
- Deployment “Self-hostable (open-weights) + hosted via AWS Bedrock + Azure AI + Together AI + Groq + Fireworks + Replicate + Databricks” → “Both (self-hosted + managed cloud) — open weights (llama.com / Hugging Face); hosted via AWS Bedrock, Azure AI, Together, Groq, Fireworks, Databricks” source
Related systems
References (2)
- Azure Machine Learning depends on at runtime — adjacent-infrastructure cell: HuggingFace Transformers; vLLM; llama.cpp; Together AI; Groq; AWS Bedrock; Azure AI Foundry
- Bedrock AgentCore (AWS) depends on at runtime — adjacent-infrastructure cell: HuggingFace Transformers; vLLM; llama.cpp; Together AI; Groq; AWS Bedrock; Azure AI Foundry
Referenced by (11)
- agentgateway integrates with — agentgateway integrations docs have a Meta (Llama API) provider page. (verified 2026-10-29)
- CrewAI integrates with — CrewAI LLM docs include a Meta Llama API provider configuration section. (verified 2026-10-29)
- Diffbot depends on at runtime — GraphRAG-fine-tuned model based on open-source Llama 3.3.
- Haystack (framework) integrates with — Haystack integrations directory has a Llama API integration page ("Use Llama Models with Haystack"). (verified 2026-10-29)
- Highlight AI depends on at runtime — adjacent-infrastructure cell: Anthropic / OpenAI APIs; local Llama option
- LiteLLM integrates with — LiteLLM docs have a Meta Llama API provider page. (verified 2026-10-29)
- LlamaIndex (framework) integrates with — LlamaIndex has an official LLM integration package llama-index-llms-meta in its repo. (verified 2026-10-29)
- OpenRouter integrates with — OpenRouter lists Meta as an inference provider (providers API slug "meta"). (verified 2026-10-29)
- Phidata (now Agno) integrates with — Agno ships a meta model provider module (agno.models.meta). (verified 2026-10-29)
- Pydantic-AI integrates with — Pydantic AI docs have a Meta (Llama API) model provider page. (verified 2026-10-29)
- Strands Agents (AWS) integrates with — Strands Agents docs have a Llama API model provider page. (verified 2026-10-29)