Llama 4 Maverick 17B 128e Instruct
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…
Llama-4-Maverick-17B-128E-Instruct-FP8Released 5 Apr 2025 (Apr 2025)Where to run it
10 hosts serve Llama 4 Maverick 17B 128e Instruct. The comparison is on blended price, 3 parts input to 1 part output.
| Host | In / out per 1M | Model id | |
|---|---|---|---|
Llama APILab API Llama-4-Maverick-17B-128E-Instruct-FP8 | - | Llama-4-Maverick-17B-128E-Instruct-FP8 | |
Lambda AIInference llama-4-maverick-17b-128e-instruct-fp8 | $0.05 / $0.10 | llama-4-maverick-17b-128e-instruct-fp8 | |
DeepInfraInference Llama-4-Maverick-17B-128E-Instruct-FP8 | $0.20 / $0.80 | Llama-4-Maverick-17B-128E-Instruct-FP8 | |
GMI CloudInference Llama-4-Maverick-17B-128E-Instruct-FP8 | $0.25 / $0.80 | Llama-4-Maverick-17B-128E-Instruct-FP8 | |
Novita AIInference llama-4-maverick-17b-128e-instruct-fp8 | $0.27 / $0.85 | llama-4-maverick-17b-128e-instruct-fp8 | |
Azure AI FoundryCloud Llama-4-Maverick-17B-128E-Instruct-FP8 | $0.25 / $1.00 Retires 19 Nov 2026 | Llama-4-Maverick-17B-128E-Instruct-FP8 | Retires 19 Nov 2026 |
Vertex AICloud llama-4-maverick-17b-128e-instruct-maas | $0.35 / $1.15 | llama-4-maverick-17b-128e-instruct-maas | |
IBM watsonx.aiCloud llama-4-maverick-17b-128e-instruct-fp8 | $0.371 / $1.48 | llama-4-maverick-17b-128e-instruct-fp8 | |
Oracle OCICloud meta.llama-4-maverick-17b-128e-instruct-fp8 | $0.72 / $0.72 | meta.llama-4-maverick-17b-128e-instruct-fp8 | |
SambaNovaInference Llama-4-Maverick-17B-128E-Instruct | $0.63 / $1.80 | Llama-4-Maverick-17B-128E-Instruct |
Capabilities
Core
- Streaming
- Function calling
- JSON mode
- Structured outputs
Multimodal
- Vision
- Image input
- Audio
- Image generation
- Video
Advanced
- Reasoning
- Tool calling
- MCP
- Prompt caching
- Batch API
- Embeddings
- Takes in
- ImagesText
- Produces
- Text
- Open weights
- License: Llama Community LicenseRuns in about 963.8 GB of GPU memoryRuntimes: Ollama, TGI, llama.cpp, vLLM
Similar models
This model has no listed price, so these are current models with similar capabilities.
Smaller Meta models
Other frontier labs
Open-weight picks
Call it
Using Llama API's API.
from litellm import completion
response = completion(
model="meta_llama/Llama-4-Maverick-17B-128E-Instruct-FP8",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Quick answers
How much does Llama 4 Maverick 17B 128e Instruct cost?
Its lowest listed price is at Lambda AI: $0.05 per million input tokens and $0.10 per million output tokens.
Which APIs offer Llama 4 Maverick 17B 128e Instruct?
Llama 4 Maverick 17B 128e Instruct runs on 10 APIs: Llama API, Lambda AI, DeepInfra, GMI Cloud, Novita AI, Azure AI Foundry, Vertex AI, IBM watsonx.ai, Oracle OCI and SambaNova.
What is Llama 4 Maverick 17B 128e Instruct's context window?
1M tokens, with up to 4K tokens in one reply.
Is Llama 4 Maverick 17B 128e Instruct being retired?
Azure AI Foundry retires it on 19 Nov 2026. 9 hosts keep serving it.
When was Llama 4 Maverick 17B 128e Instruct released?
Meta released it on 5 Apr 2025.
What is Llama 4 Maverick 17B 128e Instruct's API model id?
Use Llama-4-Maverick-17B-128E-Instruct-FP8 with Llama API. Other hosts use their own ids: llama-4-maverick-17b-128e-instruct-fp8 on Lambda AI, llama-4-maverick-17b-128e-instruct-fp8 on Novita AI, llama-4-maverick-17b-128e-instruct-maas on Vertex AI, llama-4-maverick-17b-128e-instruct-fp8 on IBM watsonx.ai.
Sources
- SambaNova pricing
- Azure AI Foundry pricing
- DeepInfra pricing
- GMI Cloud pricing
- Llama API pricing
- IBM watsonx.ai pricing
- Vertex AI pricing
- Oracle OCI pricing
- LiteLLM model map
Raw catalog entry for Llama-4-Maverick-17B-128E-Instruct-FP8 on Llama API
{
"litellm_provider": "meta_llama",
"max_input_tokens": 1000000,
"max_output_tokens": 4028,
"mode": "chat",
"source": "https://ai.developer.meta.com/docs/models",
"supported_modalities": [
"text",
"image"
],
"supported_output_modalities": [
"text"
],
"supports_function_calling": true,
"supports_tool_choice": true
}