Models · Meta
Llama 4 Scout
Llama 4 Scout is Meta's open-source Mixture-of-Experts frontier model, offering strong reasoning and coding at a fraction of the cost of closed-source alternatives. Available via Groq with blazing fast inference.
Input $0.11 per 1M tokens
Output $0.34 per 1M tokens
Context 128K tokens · $0.45 combined per 1M
Benchmark performance
Source: Artificial Analysis. Updated 2026-09-20.
Key strengths
- — Open-source frontier MoE model
- — Excellent cost-performance ratio
- — Fast inference via Groq LPUs
- — Strong general reasoning capabilities
Best use cases
- Open-source AI deployments
- Cost-sensitive production applications
- Fine-tuning for custom domains
- Research and experimentation
Pricing details
| Input (per 1M tokens) | $0.11 |
| Output (per 1M tokens) | $0.34 |
| Context window | 128K tokens |
| Architecture | 310B MoE (17B active parameters), 128K context, available on Groq's ultra-fast LPU infrastructure |
| Provider | Meta |
MoE frontier, via Groq
Recent coverage
Meta’s Muse admits it can’t explain how it reads your texts
GPT-6 Astra stabs doll in 17 of 20 RoboHarm safety tests
Gemini hacked 3 real firms in May test; Google stayed silent
Zuckerberg, Musk Block AI Regulator as Trump Calls Risks a 'Hoax'
Meta’s Muse for Mac automates local files and apps, US-only launch
The sharpest AI news, curated daily.
One email each morning. Summarized, sourced, and skimmable in under three minutes.
Pricing data is sourced from official API documentation and updated weekly. Actual costs may vary based on prompt caching, batch processing, service tiers, and regional pricing. Inblix is not affiliated with Meta.