Llama 4 Maverick

Claims and evidence

• Released 5 April 2025 ● Primary source[1]page date; Takeaways [2]Model Release Date

• Status AvailableOpen weights still offered (gated, on request) from Meta's meta-llama organisation on Hugging Face and from Meta's Llama 4 page on 2026-10-01. The hosted Llama API preview is a separate matter, see notes. ● Primary source[5]model page and access request form [6]Llama 4 Maverick, Download

• Successor of Llama 3.3 EditorialEditorial link along the Llama line: Llama 4 Maverick is among the first Llama 4 models, released after Llama 3.3, and the announcement compares its price and quality with Llama 3.3 70B. Meta does not call it the successor of a named model. ● Primary source[1]Post-training our new models
• Derived from (distillation) Llama 4 BehemothMeta describes this as codistillation, with Llama 4 Behemoth as the teacher model. ● Primary source[1]Pushing Llama to new sizes: The 2T Behemoth

• Change · Architecture One of Meta's first Llama models built on mixture of experts: 128 routed experts plus a shared expert in alternating dense and MoE layers, 17B of 400B parameters active.Compared with Llama 3.3 ● Primary source[1]
• Change · Modality Accepts images alongside text as input; Meta describes native multimodality with early fusion of text and vision tokens.Compared with Llama 3.3 ● Primary source[2]Model Information table, Input modalities [1]
• Change · Context length Context length of 1M tokens, up from 128K in Llama 3.Compared with Llama 3.3 ● Primary source[2]Model Information table, Context length [1]
• Change · Other Codistilled from Llama 4 Behemoth as teacher, then post-trained with lightweight supervised fine-tuning, online reinforcement learning and lightweight DPO.Compared with Llama 3.3 ● Primary source[1]

• Input text, imageMeta's docs give text plus up to 5 images as input; the model card says image understanding was tested up to 5 input images and the docs say image understanding is English-only. ● Primary source[2]Model Information table, Input modalities [3]Introduction, feature table, Multimodal
• Output textThe docs say text-only output; the model card table lists the output as multilingual text and code. ● Primary source[3]Introduction, feature table, Multimodal [2]Model Information table, Output modalities
• Open weights YesGated download under the Llama 4 Community License Agreement, a custom commercial license (model card); released as BF16 and FP8 weights. ● Primary source[1]opening paragraphs [5] [2]License; Quantization
• Context window 1M tokensThe announcement gives no context length for Maverick. ● Primary source[2]Model Information table, Context length [3]Introduction, feature table, Maximum Context Length
• Parameters 17B (Activated), 400B (Total)Mixture-of-experts model with 128 routed experts and a shared expert in alternating dense and MoE layers; the value is the total parameter count, of which 17B are active per token. ● Primary source[2]Model Information table, Params [3]Introduction, feature table [1]Pre-training; Post-training our new models
• Access Open-weights download, APIAPI access through Meta's Llama API, a limited free preview announced on 2025-04-29. The launch post said partner availability would follow in the coming days; cloud-partner is left out because no Meta source read here names the partners. ● Primary source[1]opening paragraphs [5] [4]Llama API

Lineage

Variants

No variants recorded in this record.

Related AI Radar coverage

AI Radar coverage starts in June 2026; no coverage linked yet.

All model releases from Meta on AI Radar →