Llama-4-Scout-17B-16E-Instruct

Model overview

Llama-4-Scout-17B-16E-Instruct is published by meta-llama as an image-text-to-text instruction model. The captured configuration identifies Llama4ForConditionalGeneration with model type llama4 and about 108.64B Safetensors parameters. According to the model card, it is a 17B-active mixture-of-experts model with 16 experts for chat and visual reasoning, under a custom Llama 4 Community License.

Recorded capabilities

Use cases in the source record

• Assistant-like chat and visual reasoning, including image recognition, captioning, and general image question answering.
• Steerable system-prompt deployments and synthetic-data or distillation workflows permitted under the publisher's community license.

Limitations and unknowns

• Benchmark figures in the model card are publisher-reported comparisons and were not independently verified.
• Provider state is historical snapshot data, including one error listing, and should be refreshed before being presented as current.

Source and provenance

Source: meta-llama/Llama-4-Scout-17B-16E-Instruct

Captured: Unknown. Processed: 2026-09-07T19:34:51.792065+00:00.

Model Information The Llama 4 collection of models are natively multimodal AI models that enable text and multimodal experiences. These models leverage a mixture-of-experts architecture to offer industry-leading performance in text and image understanding. These Llama 4 models mark the beginning of a new era for the Llama ecosystem. We are launching two efficient models in the Llama 4 series, Llama 4 Scout, a 17 billion parameter model with 16 experts, and Llama 4 Maverick, a 17 billion parameter model with 128 experts. Model developer : Meta Model Architecture: The Llama 4 models are auto-regressive language models that use a mixtu…