← CanIRunLocal · All hardware · All models
Yes. Laya 421M (Convai, decision) fits in fast memory on M2 Max (38c GPU) 96GB at BF16 (1.5 GB needed, 72.0 GB usable), about 381 tok/s (estimate).
Hardware: M2 Max (38c GPU) 96GB · Model: Laya 421M (Convai, decision)
M2 Max (38c GPU) 96GB has 96 GB of unified memory and 400 GB/s of memory bandwidth. The checker leaves macOS a share of unified memory, so Ollama can use 72.0 GB. Laya 421M (Convai, decision) (0.421B, Convai Innovations) at BF16 is a 0.84 GB download and needs 1.5 GB with headroom for about 8k tokens.
| Best quant | BF16 |
|---|---|
| Download size | 0.84 GB |
| Memory needed | 1.5 GB |
| Usable fast memory | 72.0 GB |
| Fit | fits in fast memory |
| Estimated speed | ~381 tok/s (estimate) |
| Command | ollama run laya:421m-english-bf16 |
Estimates, not benchmarks. Model sizes are ollama.com download sizes; memory needed adds headroom for ~8k tokens of context; speeds are rough figures from memory bandwidth. Data updated 2026-10-23.
A project by OnDevice Labs · llms.txt