← Main benchmarks

Ryzen AI Halo — Performance Curves

AMD Ryzen AI Halo reference system · Loading results…

Prefill and token-generation throughput across increasing context depths and system profiles.

Choose comparisons

Models and profiles

Loading…
Models
Profiles

Chart view
Prompt processing

Prefill throughput

tokens/s
Text generation

Generation throughput

tokens/s
Exact measurements

Selected results

Each point measures a 2,048-token prefill or 128-token generation after the listed starting depth. Values after ± are standard deviations.