PocketClawvol. 1 · 2026

Llama

Meta's family of open-weight large language models. Llama 3.3 70B is the leading open frontier model in 2026.

Llama 3.3 70B (Q4 quantised) is the most capable model that fits on small-form-factor hardware in 2026, achieving roughly 80% of cloud frontier performance on agentic tasks. Llama 3 8B is the small-model standard. Llama 4 is rumoured for Q3 2026.

Related terms

Local LLMOllamaQuantisation

Found a definition that's wrong, dated or could be sharper? Email us — we update with attribution unless you'd rather we didn't.