Compare
Compare local AI hardware with workload-aware output.
RTX 5090 32GB wins for coding in balanced mode
Based on model fit, speed, and quality across top recommendations.
Radeon Pro W7800 32GB
SQwen 3.6 27B
llama.cppQ4_K_MRuns well
21.5 GB / 32.0 GB
16.9 tok/s187K ctx
SDevstral Small 2 24B Instruct
llama.cppQ4_K_MRuns well
21.2 GB / 32.0 GB
25.0 tok/s87K ctx
SQwen 3.5 27B
llama.cppQ4_K_MRuns well
23.7 GB / 32.0 GB
22.3 tok/s58K ctx
Quick comparison
| Metric | Radeon Pro W7800 32GB | RTX 5090 32GB |
|---|---|---|
| Models that fit | 3 | 3 |
| Avg decode tok/s | 21.4 | 75.9 |
| Best grade score | 95 | 100 |