What could change the answer
- Long context and KV cache
- Batch size or concurrent users
- Runtime overhead and driver support
- CPU/RAM offload
- Image encoders and auxiliary models
MODEL × HARDWARECan RTX 4090 24GB
Can RTX 4090 24GB
run Llama 3.3 70B?
Only with optimization or offload.
Q4_K_M MEMORY VERDICTOFFLOAD
50.8 GB estimated requirement vs 24.0 GB usable-memory assumption.
Model weights42.9 GB
Runtime reserveIncluded
Hardware stackCUDA
Speed claimNot estimated