All Meta Llama models ranked by benchmark performance.
Unless noted otherwise, ranking surfaces on this page use BenchLM's provisional leaderboard lane rather than the stricter sourced-only verified leaderboard.
Bottom line: Meta's Llama models are open-weight and free, but benchmark coverage is still sparse for the newest entries. Llama 4 Maverick leads on coding and reasoning. Llama 4 Scout offers 10M context.
According to BenchLM.ai, Llama 3.1 405B leads this ranking with a score of 43, followed by Llama 3 70B (28) and Llama 4 Scout (24). There is a significant gap between the leading models and the rest of the field.
All models in this ranking are open-weight, meaning they can be self-hosted for maximum control and cost efficiency.
This ranking is based on provisional overall weighted scores across BenchLM.ai's scoring formula tracked by BenchLM.ai. For detailed model profiles, click any model name below. To compare two specific models head-to-head, use the "vs #" links.
Llama 3.1 405B
Meta · 128K
Best-ranked Meta model by verified scores. Most complete data.
Llama 3 70B
Meta · 128K
Llama 4 Scout
Meta · 10M
Largest context window (10M tokens). Benchmark coverage still building.
Llama 4 Maverick Meta's strongest entry for coding and reasoning tasks.
Llama 4 Scout offers the largest context window of any model at 10M tokens.
Llama 3.1 405B most complete benchmark coverage among Meta models.
Get notified when models move. One email a week with what changed and why.
Free. No spam. Unsubscribe anytime.
The top model is Llama 3.1 405B by Meta with a provisional score of 43.
The best open-weight model is Llama 3.1 405B at position #1.
5 models are included in this ranking.
Models are ranked by the same overall BenchLM score used across all leaderboards. Comparing within Meta's lineup helps identify which model fits your use case and budget.
This page only shows Meta models. Cross-provider comparison requires the overall or category-specific leaderboards. Newer models may have limited benchmark coverage initially.
For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.