Skip to content
SKRLR AI LAB
BenchmarksSecurityHardware / LabAbout
@therealskrlr

Benchmarks / Research

All cards.

Filter by model class. Each entry is markdown with a schema-checked front matter — missing caveats fail the build.

Allagentimage-genllmsystemupscalevideo-gen
DRAFTS

38 s (median)

8-step distilled · 107 frames @ 768×448

vs full 20-step final: 320 s (median)

HUNYUANVIDEO 1.5 FAST

41 s (median)

5 s clip

LOCAL LLM DECODE

~199 tok/s

Qwen3.8-Flash-Next · IQ3_S · speculative decoding

OVERNIGHT AGENT · LESSON

52 tok/s (avg)

176/176 tests passed · game blank · 0/36 tasks

SEEDVR2 UPSCALE

146 s

121 frames → 1080p

SHARED 24H · WHOLE APP

206 jobs · 0 failed

3 people · 0 s median wait · 24h

TEXT-TO-IMAGE

2-2.5 s (median)

per image · shared workflow

Real numbers from the logs.@therealskrlr· RSS · Matt@skrlr.com