Llama 4 Scout

multimodal

Llama 4 Scout is the lightest model in the next generation of the Llama 4 family, released on April 5, 2025. Llama 4 scout has a MoE architecture of 16 experts, 109 billion of which only 17 billion are permanently active. The model supports native multimodality. It is capable of simultaneously processing text and up to 5 images as input, allowing it to be used effectively in computer vision tasks such as object recognition, image description, and image-based question answering. Creative text generation, step-by-step reasoning, and programming are also among the model's strengths

That said, Llama scout can run efficiently on just a single H100 GPU, making it attractive for local deployment, and with more parameters, Llama scout outperforms lighter weight models in the segment on benchmarks.

But perhaps the most important feature of the model is its record-breaking context window support of up to 10 million tokens! which makes it ideal for working with large documents, code, multi-page queries and lengthy dialog sessions.


Announce Date: 05.04.2025
Parameters: 109B
Experts: 16
Activated at inference: 17B
Context: 10M
Layers: 48, using full attention: 12
Attention Type: Chunked Attention
VRAM requirements: 536.3 GB using 4 bits quantization
Developer: Meta AI
Transformers Version: 4.51.0
License: LLAMA 4

Public endpoint

Use our pre-built public endpoints for free to test inference and explore Llama 4 Scout capabilities. You can obtain an API access token on the token management page after registration and verification.
Model Name Context Type GPU TPS Status Link
RedHatAI/Llama-4-Scout-17B-16E-Instruct-quantized.w4a16 250,000.0 Public 2×TeslaH100 32.12 AVAILABLE chat

API access to Llama 4 Scout endpoints

curl https://chat.immers.cloud/v1/endpoints/Llama-4-Scout-17B-16E-Instruct/generate/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer USER_API_KEY" \
-d '{"model": "Llama-4-Scout-17B-16E-Instruct", "messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Say this is a test"}
], "temperature": 0, "max_tokens": 150}'
$response = Invoke-WebRequest https://chat.immers.cloud/v1/endpoints/Llama-4-Scout-17B-16E-Instruct/generate/chat/completions `
-Method POST `
-Headers @{
"Authorization" = "Bearer USER_API_KEY"
"Content-Type" = "application/json"
} `
-Body (@{
model = "Llama-4-Scout-17B-16E-Instruct"
messages = @(
@{ role = "system"; content = "You are a helpful assistant." },
@{ role = "user"; content = "Say this is a test" }
)
} | ConvertTo-Json)
($response.Content | ConvertFrom-Json).choices[0].message.content
#!pip install OpenAI --upgrade

from openai import OpenAI

client = OpenAI(
api_key="USER_API_KEY",
base_url="https://chat.immers.cloud/v1/endpoints/Llama-4-Scout-17B-16E-Instruct/generate/",
)

chat_response = client.chat.completions.create(
model="Llama-4-Scout-17B-16E-Instruct",
messages=[
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Say this is a test"},
]
)
print(chat_response.choices[0].message.content)

Private server

Rent your own physically dedicated instance with hourly or long-term monthly billing.

We recommend deploying private instances in the following scenarios:

  • maximize endpoint performance,
  • enable full context for long sequences,
  • ensure top-tier security for data processing in an isolated, dedicated environment,
  • use custom weights, such as fine-tuned models or LoRA adapters.

Recommended configurations for hosting Llama 4 Scout

Prices:
Name vCPU RAM, MB Disk, GB GPU Price, hour
teslaa2-6.32.128.160
250,000.0
32 131072 160 6 $1.65 Launch
teslaa10-4.16.128.160
250,000.0
16 131072 160 4 $1.75 Launch
rtxa5000-4.16.128.160.nvlink
250,000.0
16 131072 160 4 $2.34 Launch
teslaa100-1.16.128.160
250,000.0
16 131072 160 1 $2.71 Launch
rtx3090-4.16.96.320
250,000.0
16 98304 320 4 $3.18 Launch
teslav100-3.64.256.320
250,000.0
64 262144 320 3 $3.89 Launch
rtx4090-4.16.96.320
250,000.0
16 98304 320 4 $4.22 Launch
rtx5090-3.16.96.160
250,000.0
16 98304 160 3 $4.34 Launch
teslah100-1.16.128.160
250,000.0
16 131072 160 1 $5.23 Launch
h200-1.16.128.160
250,000.0
16 131072 160 1 $6.98 Launch
teslaa100-8.44.704.160.nvlink
10,485,760.0
44 720896 160 8 $20.40 Launch
h200-6.52.896.640
10,485,760.0
52 917504 640 6 $41.79 Launch
Prices:
Name vCPU RAM, MB Disk, GB GPU Price, hour
teslaa100-2.24.256.240
250,000.0
24 262144 240 2 $5.36 Launch
h200-1.16.128.240
250,000.0
16 131072 240 1 $6.98 Launch
rtx4090-8.44.256.240
250,000.0
44 262144 240 8 $8.59 Launch
rtx5090-6.44.256.240
250,000.0
44 262144 240 6 $8.86 Launch
teslah100-2.24.256.240
250,000.0
24 262144 240 2 $10.41 Launch
h200-6.52.896.640
10,485,760.0
52 917504 640 6 $41.79 Launch
Prices:
Name vCPU RAM, MB Disk, GB GPU Price, hour
teslaa10-1.16.32.160
250,000.0
16 32768 160 1 $0.53 Launch
teslat4-2.16.32.160
250,000.0
16 32768 160 2 $0.54 Launch
teslaa2-2.16.32.160
250,000.0
16 32768 160 2 $0.57 Launch
rtx2080ti-2.12.64.160
250,000.0
12 65536 160 2 $0.69 Launch
rtx3090-1.16.24.160
250,000.0
16 24576 160 1 $0.88 Launch
rtx4090-1.16.32.160
250,000.0
16 32768 160 1 $1.15 Launch
teslav100-1.12.64.160
250,000.0
12 65536 160 1 $1.20 Launch
rtxa5000-2.16.64.160.nvlink
250,000.0
16 65536 160 2 $1.23 Launch
rtx3080-3.16.64.160
250,000.0
16 65536 160 3 $1.43 Launch
rtx5090-1.16.64.160
250,000.0
16 65536 160 1 $1.59 Launch
teslaa100-1.16.64.160
250,000.0
16 65536 160 1 $2.58 Launch
teslah100-1.16.64.160
250,000.0
16 65536 160 1 $5.11 Launch
h200-1.16.128.160
250,000.0
16 131072 160 1 $6.98 Launch
teslaa100-8.44.512.160.nvlink
10,485,760.0
44 524288 160 8 $20.02 Launch
h200-4.32.768.480
10,485,760.0
32 786432 480 4 $28.19 Launch

Related models

Need help?

Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.