Llama 4 Maverick
Meta · 1 configuration
Intelligence10.0
Input / 1M$0.26
Output / 1M$0.91
Context1M
Tariff: Meta · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost1 measured configurations · 0 without both values
Selected: thinking default · Click a point to switch configuration
Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward...
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
10.0 pointsGPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
67.1 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
4.9 %SciCodeCode generation for scientific research problems.
31.7 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
7.9 %τ²-BenchTool use and interaction in assistance scenarios. Does not replace an evaluation of your integration.
17.8 %IFBenchVerifiable instruction following. A proxy for precision rather than a benchmark of creative style.
43.0 %MMMU-ProMultimodal understanding of academic problems that include images.
62.1 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
24.9 %Specifications from the catalog
Inputtext, image
Outputtext
Tool callingYes
Structured outputYes
Speed100.9 tok/s
First token0.94 s
Maximum output16.4K
Catalog listing date2025-04-05