Beam 1.0 501B-A23B

Running Beam 1.0 501B-A23B locally

At Q8_0 and 4K of context, Beam 1.0 501B-A23B needs about 556.9 GB — 505.9 GB of weights, 0.31 GB of KV cache and a 50.6 GB activation buffer. No card in this index clears that comfortably at Q8_0, so at that level it is a multi-GPU or CPU-offload proposition. Its smaller Q2_K build needs about 190.9 GB, which the Mac M5 Ultra 512G clears comfortably. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs.

What longer context costs

Going from 4K to 32K adds about 0.98 GB, taking the total to 558.0 GB. That is gentler than the parameter count suggests: only 9 of 52 layers keep a KV cache that grows with context, and 43 more cap theirs at a sliding window, so the cache scales at a fraction of the usual rate. The model's native window is 1M; holding all of it at Q8_0 would need about 596 GB.

Which build to download

This index only tracks Beam 1.0 501B-A23B in GGUF, across 2 levels (Q2_K, Q8_0). No per-level perplexity sweep has been published for this model, so the quality column is empty rather than estimated — within one model, more bits per weight is the only ordering the data supports. No throughput has been measured for this model on this site's hardware. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 — CUDA only, and now archived, so local use rather than serving.

Common questions

How much VRAM does Beam 1.0 501B-A23B need?

About 556.9 GB at Q8_0 with 4K of context and batch 1, rising to roughly 558.0 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.

Can Beam 1.0 501B-A23B run on a single GPU?

Not comfortably on any single card in this index at Q8_0. It needs about 556.9 GB, which means splitting across GPUs or offloading layers to system RAM. Its smaller Q2_K build needs about 190.9 GB, which the Mac M5 Ultra 512G clears comfortably.

Which quantization of Beam 1.0 501B-A23B should I use?

No published quality comparison exists for this model, so pick by footprint: Q2_K is the largest level here that fits a single machine in this index, and a higher bits-per-weight level is more faithful. All 2 levels in the index are Q2_K, Q8_0. Note this is a mixture-of-experts model — all parameters must be resident even though only a fraction are active per token, so the memory cost follows the total, not the active count.