Meta: Llama 3.3 70B Instruct — pricing & benchmarks

By meta-llama · Updated 2026-10-06 · Open weights · llama3.3cheap high-volumeopen weights / self-hostable

What Meta: Llama 3.3 70B Instruct is used for

Meta: Llama 3.3 70B Instruct is best suited for cheap high-volume, open weights / self-hostable. Access it through one Appmo API key alongside 458 other models — Appmo's optimizer can route to it automatically when it is the best value for a request. Weights: huggingface.co/meta-llama/Llama-3.3-70B-Instruct.

Use it via the Appmo API (OpenAI-compatible)

curl https://appmo.com/v1/chat/completions \
  -H "Authorization: Bearer $APPMO_KEY" \
  -d '{"model":"meta-llama/llama-3.3-70b-instruct","messages":[{"role":"user","content":"Hello"}]}'

FAQ

How much does Meta: Llama 3.3 70B Instruct cost?
Meta: Llama 3.3 70B Instruct costs $0.22 per 1M input tokens and $0.5 per 1M output tokens ($0.29 blended), as of 2026-10-06. Prices are expected to fall as competition increases.

What is Meta: Llama 3.3 70B Instruct best for?
Meta: Llama 3.3 70B Instruct is best suited for cheap high-volume, open weights / self-hostable.

How do I use Meta: Llama 3.3 70B Instruct?
Send requests to Appmo's OpenAI-compatible endpoint with model id "meta-llama/llama-3.3-70b-instruct" and one API key — no separate provider account needed.

Is Meta: Llama 3.3 70B Instruct open source?
Yes — its weights are public on Hugging Face (meta-llama/Llama-3.3-70B-Instruct) under the llama3.3 license; check the license terms before commercial use.

Can Meta: Llama 3.3 70B Instruct create images?
No, Meta: Llama 3.3 70B Instruct works with text.