{"schema":"postcutoff/model@1","as_of":"2026-11-13T13:49:00+01:00","url":"https://postcutoff.com/m/llama-4-maverick/","md":"https://postcutoff.com/m/llama-4-maverick/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"llama-4-maverick","name":"Llama 4 Maverick (17B-128E)","org":"Meta","family":"Llama 4","released":"2025-04-05","status":"legacy","type":"multimodal","modality_in":["text","image"],"modality_out":["text"],"open_weights":true,"model_license":"llama4-community","context_window":1000000,"max_output":null,"knowledge_cutoff":"2024-08","pricing":null,"price_line":"Price not published","access":[{"provider":"Hugging Face","url":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct"},{"provider":"AWS Bedrock","model_id":"meta.llama4-maverick-17b-instruct-v1:0","docs":"https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-4-maverick-17b-instruct.html"},{"provider":"OpenRouter","model_id":"meta-llama/llama-4-maverick","url":"https://openrouter.ai/meta-llama/llama-4-maverick"},{"provider":"Web app","url":"https://meta.ai"}],"capabilities":[{"name":"First natively multimodal Llama (early fusion)","detail":"Llama 4 were the first Llama models with native multimodality via early fusion of text and vision tokens.","first":true,"discovered":"launch","source":"https://ai.meta.com/blog/llama-4-multimodal-intelligence/"},{"name":"400B-total MoE on one H100 host","detail":"17B active / 128 experts / ~400B total; runs on a single H100 host.","first":false,"discovered":"launch","source":"https://ai.meta.com/blog/llama-4-multimodal-intelligence/"},{"name":"LMArena experimental-variant controversy","detail":"Launch LMArena Elo 1417 came from an experimental chat-tuned variant, not the released weights, drawing criticism.","first":false,"discovered":"later","source":"https://en.wikipedia.org/wiki/Llama_(language_model)"}],"entry":null,"notes":"FP8 repo meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8. Bedrock max output 8K. No first-party pay-as-you-go pricing verified. Superseded at Meta by closed Muse Spark and open Muse Glimmer.","verified":"2026-09-29","body_md":"Open-weight MoE multimodal model; still widely hosted and cheap on third-party providers.\n\n```bash\ncurl https://openrouter.ai/api/v1/chat/completions -H \"Authorization: Bearer $OPENROUTER_API_KEY\" -H \"Content-Type: application/json\" \\\n -d '{\"model\":\"meta-llama/llama-4-maverick\",\"messages\":[{\"role\":\"user\",\"content\":\"Hello\"}]}'\n```\n\nSources: https://ai.meta.com/blog/llama-4-multimodal-intelligence/ , https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-meta-llama-4-maverick-17b-instruct.html","page_url":"https://postcutoff.com/m/llama-4-maverick/","events_after":2056,"major_after":487,"historic_after":97,"missing_at_launch":29,"events_since_release":null,"you_url":"https://postcutoff.com/you/llama-4-maverick/","briefings":{"xs":"https://postcutoff.com/briefings/cutoff-2024-08-xs.md","s":"https://postcutoff.com/briefings/cutoff-2024-08-s.md","m":"https://postcutoff.com/briefings/cutoff-2024-08-m.md","full":"https://postcutoff.com/briefings/cutoff-2024-08.md"}}