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

nemotron-3.5-lightning-30b-a3b leads 6 rows, nemotron-3.5-bolt-e4b leads 1.

nemotron-3.5-lightning-30b-a3bMetricnemotron-3.5-bolt-e4b
nvidia ServingStatusServing nvidia
96●Score95
100%Uptime100%
190 msFirst token (median)●118 ms
1.38 s●Full reply (median)1.62 s
1.13 s●Best reply1.46 s
163 tok/s●Throughput (median)83.8 tok/s
8●Round wins1
#1●Leaderboard rank#2
75.0%Faster when both answered (12 rounds)25.0%

marks the better value in each row

Full reply time overlay

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

nemotron-3.5-lightning-30b-a3bnemotron-3.5-bolt-e4b no answer
0 ms2.50 s5.00 s7.50 s10.0 s11-17 10:4311-17 16:13 UTC

Win timeline

Per round, oldest left: which model replied faster.

Anemotron-3.5-lightning-30b-a3b Bnemotron-3.5-bolt-e4b ·neither answered