Skip to main content

Includes 694 hourly rounds up to 2026-11-16 from NIMStats by Mauro Druwel (source), measured with a different prompt and timeout. Our own probe adds a round every 30 minutes.

Model comparison

Head-to-head from live probes, last 24 hours.

⇄

Head-to-head stats

gpt-oss-20b leads 0 rows, step-5-preview leads 7.

gpt-oss-20bMetricstep-5-preview
openai DownStatusUnstable stepfun
0Score●45
0%Uptime●50.0%
n/aFirst token (median)●7.49 s
n/aFull reply (median)●25.4 s
n/aBest reply●7.40 s
n/aThroughput (median)●10.7 tok/s
0Round wins0
#14Leaderboard rank●#10
n/aFaster when both answered (0 rounds)n/a

marks the better value in each row

Full reply time overlay

Each probe, both models on one axis. Lower is better.

gpt-oss-20bstep-5-preview no answer
0 ms50.0 s100.0 s150.0 s200.0 s11-17 10:4311-17 16:13 UTC

Win timeline

Per round, oldest left: which model replied faster.

Agpt-oss-20b Bstep-5-preview ·neither answered