LLaMA 4 Scout is a 17-billion parameter model leveraging a Mixture-of-Experts architecture with 16 active experts, positioning it as the top multimodal model in its category. It consistently outperforms competitors like Gemma 3, Gemini 2.0 Flash-Lite, and Mistral 3.1 across diverse benchmark tasks. Despite its performance, LLaMA 4 Scout is remarkably efficient—capable of running on a single NVIDIA H100 GPU with Int4 quantization. It also boasts an industry-leading 10 million token context window and is natively multimodal, seamlessly processing text, images, and video inputs for advanced real-world applications.
Amazon Nova Micro is a text-only model optimized for cost and speed. With a context window of 128K tokens, it excels at tasks like text summarization, translation, interactive chat, and basic coding. Released as part of the Amazon Nova foundation models, it supports fine-tuning and distillation for customization on proprietary data.
| Llama 4 Scout | Nova Micro | |
|---|---|---|
Web Site
| ||
Provider
| ||
Chat
| ||
Release Date
| ||
Modalities
| text images video | text |
API Providers
| Meta AI, Hugging Face, Fireworks, Together, DeepInfra | Amazon Bedrock |
Knowledge Cut-off Date
| 2025-04 | Purposefully not disclosed |
Open Source
| Yes (Source) | No |
Pricing Input
| Not available | $0.04 per million tokens |
Pricing Output
| Not available | $0.14 per million tokens |
MMLU
| Not available | 77.6% CoT Source |
MMLU-Pro
| 74.3% Reasoning & Knowledge Source | - |
MMMU
| 69.4% Image Reasoning Source | - |
HellaSwag
| Not available | - |
HumanEval
| Not available | 81.1% pass@1 Source |
MATH
| Not available | 69.3% CoT Source |
GPQA
| 57.2% Diamond Source | 40% Main Source |
IFEval
| Not available | 87.2% Source |
SimpleQA
| - | - |
AIME 2024 | - | - |
AIME 2025 | - | - |
Aider Polyglot
| - | - |
LiveCodeBench v5
| - | - |
Global MMLU (Lite)
| - | - |
MathVista
| - | - |
Mobile Application | - | - |
Compare AI. Test. Benchmarks. Mobile Apps Chatbots, Sketch
Copyright © 2025 All Right Reserved.