Meta (via Together.ai) API Pricing — April 2026
Meta's Llama models are available open-weight via multiple hosting providers, with pricing set by the infrastructure layer rather than the model creator.
Model Pricing
| Model | Input ($/M tokens) | Output ($/M tokens) | Context | Use Cases |
|---|---|---|---|---|
| Llama 3.3 70B | $0.88 | $0.88 | 128K | General-purpose production workloads |
| Llama 3.3 70B (Fireworks) | $0.90 | $0.90 | 128K | General-purpose production workloads |
| Llama 3.1 70B | $0.88 | $0.88 | 128K | General-purpose production workloads |
| Llama 3.2 90B Vision unverified | $0.88 | $0.88 | 128K | General-purpose production workloads |
| Llama 3.2 11B Vision unverified | $0.18 | $0.18 | 128K | General-purpose production workloads |
How Meta (via Together.ai) compares to other providers
OpenAI PricingAnthropic PricingGoogle PricingMistral PricingCohere PricingxAI PricingDeepSeek PricingTogether.ai PricingFireworks.ai PricingGroq PricingAmazon Bedrock PricingCerebras PricingNVIDIA NIM Pricing
Running Meta (via Together.ai) in production? Clawback audits your spend and shows you exactly where you can cut costs without sacrificing quality.
Audit your Meta (via Together.ai) spend →