Inference
EN
Meta

Llama 4 Maverick

Llama 4 Maverick is a standard model from Meta. 15 benchmarks count toward its index, in 7 categories.

  • Open weights
  • Standard
  • Input: text · image
  • Released 2025-04
RankedEnough results for an index.
IndexAll categories, out of 100.34.4±4.7Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.61/sMedian 67/s
First tokenSeconds to the first output token.0.6 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.35Median $0.60
Output / 1MUS dollars per 1M output tokens.$1.00Median $2.55
ContextMaximum tokens in one request.524KMedian 262K
EloLMArena rating and rank.—Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
meta-llama/llama-4-maverick
Max outputMaximum output tokens per request.
33K
Knowledge cutoffLast date of training data.
31 Aug 2024
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.17
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
8 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Oracle CloudCloud$0.72$0.72512K
Deep InfraDeprecatedHost$0.20$0.801.05M
IO.NETHost$0.269$0.826430K
Novita AIHost$0.27$0.851.05M
DigitalOceanCloud$0.25$0.87128K
Amazon BedrockDeprecatedCloud$0.24$0.971M
OpenRouterRouter$0.35$1.00524K
IBM watsonxCloud$0.371$1.48131K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

401.58B parameters
16-bit
844.3 GB
8-bit
449.0 GB
4-bit
256.6 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.3 counted36.9184 / 218
CodingCode writing and repair.1 counted37.2—
ReasoningLogic problems and puzzles.4 counted36.7181 / 227
MultimodalTasks with images and text.1 counted36.0—
KnowledgeFacts and expert knowledge.2 counted34.0230 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted29.2—
MathMath problems.3 counted28.6146 / 168

Results

15 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
MATH level 5Math73.0% ±1.132 / 88
MazeBench: visual spatial reasoning in a 3D open worldAgentic0.0%23 / 48
GPQA diamond · Epoch AIKnowledge67.0% ±2.8117 / 217
Artificial Analysis IFBenchInstruction43.0%94 / 146
Artificial Analysis GPQA DiamondKnowledge67.1%139 / 199
Critical Physics TasksReasoning0.0%150 / 214
Artificial Analysis Long Context ReasoningReasoning50.0%154 / 217
OTIS Mock AIME 2024-2025Math20.6% ±4.9146 / 200
GDPval-AA normalizedAgentic0.0%101 / 133
Artificial Analysis MMMU-ProMultimodal62.1%88 / 113
Artificial Analysis Humanity's Last ExamKnowledge4.9%180 / 222
FrontierMath-2025-02-28-PrivateMath0.7% ±0.562 / 69
Artificial Analysis SciCodeCoding31.7%103 / 114
τ²-Bench Tool-Agent-User EvaluationAgentic17.8%136 / 149
ARC-AGI-2 (semi-private)Reasoning0.0%120 / 131
ARC-AGI-1 (semi-private)Reasoning4.4%123 / 126
FrontierMath-2025-02-28-PublicMath0.0%33 / 64
ARC-AGI-2 (public eval)Reasoning0.0%102 / 122
ARC-AGI-1 (public eval)Reasoning7.1%114 / 116

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYARC Prize FoundationTerms restrict reusemodels.devMIT

History

5 releases
Index35.6 → 34.4
23 Oct 202612 Nov 2026
Output / 1M$0.653 → $1.00
23 Oct 202612 Nov 2026

Questions

How much does Llama 4 Maverick cost?
$0.35 per 1M input tokens, $1.00 per 1M output tokens.
How good is Llama 4 Maverick?
Its index is 34.4 out of 100. That is place 237 of 283 ranked models.
What is Llama 4 Maverick best at?
Reasoning: 36.7 out of 100, place 181 of 227.
How fast is Llama 4 Maverick?
61 output tokens per second, 0.6 s to the first token.
What is the context window of Llama 4 Maverick?
524K tokens.
Is Llama 4 Maverick open source?
Yes. Its weights are published.
When was Llama 4 Maverick released?
April 2025.

Badge

Llama 4 Maverick on Inference360: place 237 of 283, index 34.4
HTML
<a href="https://inference360.ai/models/llama-4-maverick"><img src="https://inference360.ai/badges/llama-4-maverick.svg" alt="Llama 4 Maverick on Inference360: place 237 of 283, index 34.4" height="20"></a>
Markdown
[![Llama 4 Maverick on Inference360: place 237 of 283, index 34.4](https://inference360.ai/badges/llama-4-maverick.svg)](https://inference360.ai/models/llama-4-maverick)

Compare with

GPT-4.1 mini34.5Mistral Large 334.3Gemini 2.5 Flash35.0Magistral Medium33.7Llama 3.3 70B28.2

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Meta

All models
Llama 4 Scout26.8Llama Guard 4 12BUnrankedLlama 3.3 70B InstructUnrankedLlama 3.3 70B28.2Llama 3.2 1B InstructUnranked