Llama 4 Scout pricing & benchmarks
Released Apr 5, 2025 · served by 4 providers
What Llama 4 Scout is
Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B. It supports native multimodal input...
- Model ID:
meta-llama/llama-4-scout - Modality: text+image->text
- Knowledge cutoff: 2024-08-31
- Tool calling: supported · Structured output: supported
- Weights:
meta-llama/Llama-4-Scout-17B-16E-Instructon Hugging Face
Providers and prices for Llama 4 Scout
The same model costs different amounts depending on who serves it. Prices are per 1M tokens, cheapest first. Uptime and throughput are OpenRouter's rolling measurements, not vendor claims.
| Provider | Input | Output | Context | Quant | Uptime 24h | Throughput |
|---|---|---|---|---|---|---|
| DeepInfracheapest | $0.100 | $0.300 | 328K | fp8 | 99.8% | — |
| Groq | $0.110 | $0.340 | 131K | — | 99.8% | — |
| Novita | $0.180 | $0.590 | 131K | bf16 | 99.8% | — |
| $0.250 | $0.700 | 1.3M | — | 99.9% | — |
Llama 4 Scout benchmark results
Design Arena head-to-head results, as reported through the OpenRouter model API.
| Category | Arena | Rank | Elo | Win rate |
|---|---|---|---|---|
| dataviz | models | #106 | 925 | 39.3% |
| uicomponent | models | #108 | 805 | 25.5% |
| gamedev | models | #111 | 829 | 27.4% |
| codecategories | models | #114 | 820 | 26.6% |
| website | models | #120 | 775 | 22.7% |
Cheaper models in the same class
Models scoring within 4 points of Llama 4 Scout on the Intelligence Index, but with a lower output price.
FAQ
How much does the Llama 4 Scout API cost?
$0.100 per 1M input tokens and $0.300 per 1M output tokens. A typical workload of 1M input + 300K output tokens costs about $0.19.
Is Llama 4 Scout free?
No. It is a paid model starting at $0.100 per 1M input tokens, though some providers offer trial credits.
Which provider is cheapest for Llama 4 Scout?
DeepInfra at $0.100 input / $0.300 output per 1M tokens (fp8 quantization).
Can I self-host Llama 4 Scout?
Yes — weights are published as meta-llama/Llama-4-Scout-17B-16E-Instruct on Hugging Face, so you can run it on your own hardware.