Compare

Put two chips, runtimes, or runtime versions side by side. Everything else is held fixed or called out, so a ratio is worth exactly what the matched runs say.

Add your benchmark

Runs compared: 41
Like-for-like pairs: 0
Runtime available: 1

Compare

Side A

Side B

Model

Chip

Quantisation A

Quantisation B

Backend

Prefill size

Decode size

Cooldown

41 runs vs 0 runs · 0 like-for-like pairs

Held fixedChip: AMD Radeon RX 7900 XT (RADV NAVI31)Prefill: PP512Decode: TG128

Only one runtime has matching runs

llama.cpp is the only runtime with submitted reports. A runtime comparison needs reports from a second runtime; the CLI's llama.cpp adapter is the quickest way to add one.