Skip to content
SLM-Works

Llama 3.3 (70B Instruct)

Llama · ~70B parameters · 13B+ · Last reviewed 2026-03-20

Why we use it

High-quality teacher and benchmark anchor for enterprises that need frontier-like behavior inside a private boundary before compressing to SLMs.

License summary

Meta Llama license; large checkpoints have higher infra and compliance overhead - ensure your deployment matches license redistribution rules.

Typical deployment profiles

  • VPC, multi-GPU
  • Datacenter / large clusters

Focus tags

  • General
  • Reasoning
  • Multilingual

Typical use cases

  • Teacher for distillation
  • Private “full-size” assistant
  • Difficult eval sets

External references

Related SLM-Works services

← Back to foundation models catalog · Insights · Glossary

Contact to validate a model choice for your environment.