About Llama 4 Maverick
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...
Specifications
- Provider
- Meta
- Context Length
- 1,048,576 tokens
- Input Types
- text, image
- Output Types
- text
- Category
- meta-llama/llama-4-maverick-17b-128e-instruct
- Added
- 4/5/2025
Benchmark Performance
How Llama 4 Maverick compares to its closest rivals across industry benchmarks
Evaluates AI ability to resolve real GitHub issues from Python repositories
Metric: % Resolved
#21
gpt-oss-120b
26.0
#25
Gemini 2.0 flash
13.5
#27
Qwen2.5-Coder 32B Instruct
9.0
Comprehensive comparison of AI models across intelligence, price, speed, and latency
Metric: Intelligence Index
#97
Solar Pro 2
38.0
#98
NVIDIA Nemotron Nano 9B V2
37.0
#99
Qwen3 30B A3B 2507
37.0
#100
Devstral 2
36.0
#101
NVIDIA Nemotron Nano 9B V2
36.0
#103
Nova 2.0 Lite
36.0
#104
Llama 3.3 Nemotron Super 49B
35.0
#105
Mistral Medium 3.1
35.0
#106
Nova 2.0 Omni
34.0
Frequently Asked Questions
Common questions about Llama 4 Maverick