RUN THIS LLM
Search local LLM hardware requirements
CodeLlama 34B
Meta · 34B · Code
Large code model excelling at code review, refactoring, and multi-file understanding.
VRAM Requirements
| Quantization | VRAM |
|---|---|
| Q4_K_M (smallest) | 20.4 GB |
| Q8_0 (balanced) | 37.4 GB |
| FP16 (full quality) | 68 GB |
Specifications
- Parameters: 34B
- Category: Code
- Max context: 16K tokens
- System RAM: 32 GB minimum
- HuggingFace: codellama/CodeLlama-34b-Instruct-hf
Benchmarks
- HumanEval: 48 — Code generation (pass@1)
Loading interactive analysis...