What GPU do you need to run Beam 1.0 501B A23B?
501.27B params 1M max context Tool useReasoning
The smallest GPU memory that runs Beam 1.0 501B A23B well is 512 GB (Q8_0, e.g. M3 Ultra). 2 of 132 GPUs run it well. It's a mixture-of-experts model, so all weights must fit in memory but speed depends only on the active experts.
SWE-bench Verified 81.6
VRAM needed per quantization
| Quantization | Weights | 16K context | 32K context | Full 1M |
|---|---|---|---|---|
| Q4_K_M | 292.9 GB | 293.5 GB | 294.0 GB | 327.7 GB |
| Q8_0 | 496.2 GB | 496.7 GB | 497.3 GB | 531.0 GB |
Weights plus fp16 KV cache. Leave ~1 GB headroom for the inference engine.
GPUs that can run Beam 1.0 501B A23B
Best quantization that fits on each GPU, with the context and speed it gets. For chips with several memory sizes, the smallest that works is shown.