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.

nvidia

nemotron-3-ultra-550b-a55b

Serving

Rank 6 of 15 Last probed (9 min ago)

nvidia/nemotron-3-ultra-550b-a55b answered 100% of probes on the NVIDIA NIM free API in the last 24 hours, with a median time to first token of 414 ms and 40.1 tok/s. It ranks 6 of 15 tracked models.

100% Uptime 12/12 probes
89/ 100 Score 0 round wins
414 ms First token p95 1.59 s
4.87 s Full reply median
2.50 s Best reply fastest in range
40.1tok/s Throughput median

Model vs. median

Against the median of models that answered, last 24 hours.

Uptime
100% 100%
at the median
First token
414 ms 414 ms
at the median
Full reply
4.87 s 4.91 s
1.0× better than median
Throughput
40.1 tok/s 31.0 tok/s
1.3× better than median

nemotron-3-ultra-550b-a55b median of answering models

Error breakdown

0 failed of 12 probes, last 24 hours.

No failed probes in this range.

Time to first token

Each probe. Lower is better.

0 ms500 ms1.00 s1.50 s2.00 s11-17 10:4311-17 16:13 UTC
answered no answer

Throughput

Tokens per second after the first token, each probe. Higher is better.

0.0 tok/s25.0 tok/s50.0 tok/s75.0 tok/s100 tok/s11-17 10:4311-17 16:13 UTC
answered no answer

Availability heatmap

Last 7 days by UTC hour.

all answered some failed none answered not probed

Run history

Last 20 probes.

TimeResultFirst tokenFull replySpeed
OK782 ms4.79 s50.0 tok/s
OK546 ms6.45 s33.0 tok/s
OK382 ms4.34 s50.6 tok/s
OK405 ms5.39 s40.1 tok/s
OK877 ms6.50 s35.6 tok/s
OK406 ms4.13 s53.6 tok/s
OK1.59 s4.87 s60.9 tok/s
OK397 ms2.50 s95.1 tok/s
OK461 ms6.72 s32.0 tok/s
OK414 ms7.04 s30.2 tok/s
OK369 ms13.7 s15.0 tok/s
OK717 ms3.83 s64.3 tok/s

Call nemotron-3-ultra-550b-a55b

The endpoint is OpenAI-compatible.

Create a free key at build.nvidia.com, then:

curl https://integrate.api.nvidia.com/v1/chat/completions \
  -H "Authorization: Bearer $NVIDIA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nvidia/nemotron-3-ultra-550b-a55b",
    "messages": [{"role": "user", "content": "Hello"}],
    "max_tokens": 256,
    "stream": true
  }'

Other models answering

Top of the leaderboard right now.