Meta's long-context open-weight model. 10M token context window. 109B/17B active MoE.
Each score is anchored to human baseline (100). Source URLs link to the original benchmark, leaderboard, or release note.
| Sub-capability | Quality | Source | Score vs. baseline | Score |
|---|---|---|---|---|
| Language Understanding | vendor reported | ai.meta.com/llama-4 | 90% |