Home/Models/llama-4-scout-90b

llama-4-scout-90b

MetaTextOpen weights · Llama 4 Community

Use with your agent

Context

256K

Input

$0.40/M

Output

$0.60/M

Throughput

120tok/s

Latency

300ms

Popularity

58%

Overview

Llama 4 Scout 90B is the open-weight workhorse. Run it hosted on Claws or pull the weights onto your own GPUs — either way the behaviour is identical, so you keep full portability and zero vendor lock-in.

Capabilities

LONG CONTEXTHandles very large documents and histories.

Routing & endpoints

Endpoint	Context	Price	Latency	Status
claws-edge · us-east	256K	$0.40 / $0.60 /M	300ms	Operational
claws-edge · eu-west	256K	$0.40 / $0.60 /M	322ms	Operational
Meta · direct	256K	$0.40 / $0.60 /M	348ms	Operational

Sample request

curl https://api.claws.io/v1/chat/completions \
  -H "Authorization: Bearer $CLAWS_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama-4-scout-90b",
    "messages": [{"role": "user", "content": "Hello"}]
  }'