🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Apple M1 Max 32GB run Transformer (Adaptive Input Embeddings) WT103?
Yes — here's how fast and up to what context, with real measured numbers.
Transformer (Adaptive Input Embeddings) WT103 on the Apple M1 Max 32GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 409k | 550.6-664 t/s | ~13.8-16.6 (slow) |
| Q8 | yes | 413k | 1101.2-1327.9 t/s | ~27.5-33.2 (slow) |
| Q4 | yes | 414k | 1966.5-2371.3 t/s | ~49.2-59.3 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
→ All models on the Apple M1 Max 32GB
→ Transformer (Adaptive Input Embeddings) WT103 specs
→ Other hardware
FAQ
Can the Apple M1 Max 32GB run Transformer (Adaptive Input Embeddings) WT103?▶
Yes. At Q4 it generates ~1966.5-2371.3 tok/s and fits up to 414k context; beyond that it spills to SSD swap and slows to ~49.2-59.3 tok/s.
How much context fits?▶
At Q4, up to about 414k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.