Llama 4 Maverick
Apr 2025 · Open weight · Current · supported evidence
The Llama you can actually fine-tune. Not on the 2026 intelligence frontier. Still the default open Meta line.
Intelligence
32
AA Index
API price
$0.20 / $0.60
Input / output per 1M
Context
1M
120 tok/s
Capability
Intelligence Index32
Coding index45
Agentic index18
Arena Elo (offset)110
Use it when
- ▸Fine-tunes
- ▸Permissive-ish Meta license
- ▸Local multimodal
Skip if
- –You wanted Muse Spark quality with Llama on the box
Benchmarks
GPQA Diamond Graduate-level science questions designed so Google search is not enough. Still one of the cleanest knowledge/reasoning splits. | — |
|---|---|
Humanity’s Last Exam Expert-written questions across fields. Harder than MMLU; the current differentiator for “does this model actually know things.” | — |
MMLU-Pro Harder, less-saturated successor to MMLU. Classic MMLU is above 90% for every flagship and no longer ranks the field. | 88% |
SWE-bench Verified 500 human-validated GitHub issues. Score swings 5–15 points by harness — treat vendor numbers as an upper bound. | — |
SWE-bench Pro Harder, contamination-resistant coding eval. Currently the best public split between “can code” and “can maintain a repo.” | — |
Terminal-Bench 2.1 End-to-end tasks in a real terminal. Better proxy for coding agents than HumanEval, which is fully saturated. | — |
ARC-AGI-2 Abstract visual puzzles. Rewards generalization over memorization. GPT-5.6 Sol currently leads the published set. | — |
AIME 2025 American Invitational Mathematics Examination. Contest math; reasoning-mode models dominate. | — |
MMMU College-level multimodal understanding across diagrams, charts, and exam figures. | — |
Modalities
text · vision · tools
17B active MoE
Reasoning mode
standard
Hybrid and reasoning models spend tokens thinking. That raises GPQA and agents, and also raises latency and bill.