ZMIME
Meta · Released Apr 5, 2025

Llama 4 Maverick 17B Instruct

Open multimodal Llama for strong reasoning with efficient everyday serving.

  • Open weights
  • Tool calling
  • Vision
Capability
132.2ECI · #134 of 158
Input
$0.27per 1M tokens
Output
$0.86per 1M tokens
Context
1M16K max output
Pricing

What it costs

Input
$0.27per million tokens
Output
$0.86per million tokens
Blended (3:1)
$0.417cheaper than 64% of priced models

Typical monthly bills

Monthly usageEstimated cost
Side project2M input + 0.5M output tokens$0.97
Team assistant25M input + 5M output tokens$11.05
Production app250M input + 50M output tokens$110.50

Median of 16 paid API providers on models.dev.

Capability

Independent benchmarks

132.2Epoch Capabilities Index
#134 of 158 scored models
68Median 147174

The shaded band is Epoch AI’s confidence range (130.0–134.3); the tick marks the median scored model.

  • GPQA DiamondGraduate-level science questions67.0%
  • OTIS Mock AIME 2024–2025Competition mathematics20.6%

Scores from Epoch AI, run independently of Meta.

Specifications

The details

Lab
Meta
Released
Apr 5, 2025
Knowledge cutoff
Aug 2024
Context window
1,000,000 tokens
Max output
16,384 tokens
Inputs
Text, Images
Output
Text
Reasoning
No
Tool calling
Yes
Structured output
No
Availability
16 API providerslisted on models.dev
Lab-reported

What Meta claims

Published by the lab at launch. Settings vary, so compare these only with care.

BenchmarkScoreSettingSource
Aider Polyglot15.6 percent correct—Source
SWE-Bench Pro5.24 resolve rate—Source
Same lab

More from Meta

Questions

About Llama 4 Maverick 17B Instruct

How much does Llama 4 Maverick 17B Instruct cost?

Llama 4 Maverick 17B Instruct costs $0.27 input / $0.86 output per million tokens (median across 16 API providers). At a 3:1 input-to-output mix that is $0.417 per million tokens, cheaper than 64% of the 469 priced models we track.

What is the context window of Llama 4 Maverick 17B Instruct?

Llama 4 Maverick 17B Instruct accepts up to 1,000,000 tokens per request and can write up to 16,384 tokens in one response.

How good is Llama 4 Maverick 17B Instruct?

Epoch AI gives Llama 4 Maverick 17B Instruct a Capabilities Index score of 132.2 (likely range 130.0–134.3), ranking it #134 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 67.0%, OTIS Mock AIME 2024–2025 20.6%.

Is Llama 4 Maverick 17B Instruct open source?

Yes. Meta publishes the weights, so it can be downloaded and self-hosted.

What inputs does Llama 4 Maverick 17B Instruct support?

Llama 4 Maverick 17B Instruct accepts text and images and replies in text. It is not a dedicated reasoning model, supports tool calling.

When was Llama 4 Maverick 17B Instruct released?

Meta released Llama 4 Maverick 17B Instruct on Apr 5, 2025. Its training data runs to Aug 2024.

What are the best alternatives to Llama 4 Maverick 17B Instruct?

The closest current models from other labs on capability, price and release date are DeepSeek-V3 (DeepSeek, ECI 131.7, $0.32 / $1.10), GPT-4.1 mini (OpenAI, ECI 134.3, $0.40 / $1.60), Qwen3 30B A3B (Alibaba (Qwen), ECI 134.2, $0.109 / $0.50) and Gemini 2.5 Flash-Lite Preview 06-17 (Google, ECI 133.1, $0.09 / $0.36).