Meta · language model
Llama 4 Scout: pricing, context and benchmarks
Llama 4 Scout is Meta's model released on April 5, 2025. It costs $0.1 per million input tokens and $0.3 per million output tokens, with a 1.3M-token context window. Epoch AI ranks it #117 of 125 rated models.
What you need to know about Llama 4 Scout
Meta, based in United States, put Llama 4 Scout online on April 5, 2025. It accepts text and images as input and produces text. Its context window reaches 1,310,720 tokens, larger than 99 % of the models in the registry, and it can write up to 16,384 tokens in one go.
Released more than 1 year ago, it belongs to a generation most vendors have since superseded.
Llama 4 Scout pricing
On OpenRouter, Llama 4 Scout costs $0.1 per million input tokens and $0.3 per million output tokens. Counting three tokens read for every token written, its average price is $0.15 per million, more expensive than 17 % of paid models (5.1 times cheaper than the median of $0.76).
A concrete example: a task that sends 2,000 tokens (about 1,500 words) and receives a 500-token answer costs $0.0004. At 1,000 tasks a day, budget $11 a month. The price calculator redoes this with your own volumes.
Performance and benchmarks
The Epoch Capabilities Index (ECI), which aggregates published results across many benchmarks, gives Llama 4 Scout a score of 129.6 (range 125.1 to 131.2). That places it #117 of 125 rated models.
For at least equal performance, DeepSeek V4 Flash 0731, Qwen3.7 Flash, and gpt-oss-20b cost at least 30% less.
Features, privacy and local use
- It supports tool calling, so it can drive an AI agent, call APIs or MCP servers.
- It can answer in an enforced JSON format (structured outputs).
- Its knowledge stops at August 2024.
Its weights are open and downloadable from Hugging Face: you can run it on your own machines without sending your data to a third party.
Other Meta models
| Model | Released | Avg. price | ECI |
|---|---|---|---|
| Llama Guard 4 12B | Apr 30, 2025 | $0.18 | — |
| Llama 4 Maverick | Apr 5, 2025 | $0.3 | 132.2 |
| Llama 3.3 70B Instruct | Dec 6, 2024 | $0.16 | — |
| Llama 3.2 1B Instruct | Sep 25, 2024 | $0.07 | — |
| Llama 3.2 3B Instruct | Sep 25, 2024 | $0.12 | — |
| Llama 3.1 70B Instruct | Jul 23, 2024 | $0.4 | — |
| Llama 3.1 8B Instruct | Jul 23, 2024 | $0.06 | — |
Verdict
Computed score: 55/100, from public data (performance, price, context, features).
Strengths
- Low price: cheaper than 83 % of paid models.
- Very large context (1.3M tokens): long documents or a whole codebase fit in one request.
- Open weights: self-hostable, your data stays with you.
- Agent-ready: tool calling and structured JSON output.
Limitations
- Modest performance: behind 93 % of rated models.
- Older generation, often replaced by a newer model from the same vendor.
Alternatives
- GPT-4.1 Nano$0.18
- GPT-5 Nano$0.14
- Gemma 3 27B$0.15
- Qwen3.5-9B$0.19
- GPT-4o-mini$0.26
- GPT-4o-mini (2024-07-18)$0.26
Comparisons and tools
Learn more
Frequently asked questions
How much does Llama 4 Scout cost?
$0.1 per million input tokens and $0.3 per million output tokens on OpenRouter (checked October 25, 2026). A task with 2,000 tokens read and 500 written costs $0.0004.
How big is Llama 4 Scout's context window?
1,310,720 tokens, roughly 983,000 words, with a maximum of 16,384 output tokens.
Can Llama 4 Scout run locally?
Yes, its weights are published.
Is Llama 4 Scout suitable for an AI agent?
Technically yes: it supports tool calling and structured outputs. Then check its cost for your volume.
How good is Llama 4 Scout?
Epoch AI ranks it #117 of 125 rated models (ECI 129.6).
What are the alternatives to Llama 4 Scout?
The closest models on price and performance are GPT-4.1 Nano, GPT-5 Nano, and Gemma 3 27B.
Related articles
- Mistral AI: models, pricing and agentsOctober 25, 2026
- LLM benchmarks: how to read model rankingsOctober 24, 2026
- Running a local LLM: hardware, software, and limitationsOctober 24, 2026
- Open-source or closed models for your AI agents?October 5, 2026
Neutrality. No affiliate links. No vendor pays to be reviewed, rated or ranked. Reviews are directed by François Chollet, neutral and factual, and every data point is dated and sourced. How we score