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.

stepfun

step-5-preview

Unstable

Rank 7 of 15 Last probed (22 min ago)

stepfun-ai/step-5-preview answered 74.7% of probes on the NVIDIA NIM free API in the last 7 days, with a median time to first token of 14.8 s and 22.3 tok/s. It ranks 7 of 15 tracked models.

74.7% Uptime 118/158 probes
60/ 100 Score 8 round wins
14.8 s First token p95 216.8 s
20.3 s Full reply median
1.43 s Best reply fastest in range
22.3tok/s Throughput median

Model vs. median

Against the median of models that answered, last 7 days.

Uptime
74.7% 74.7%
at the median
First token
14.8 s 14.8 s
at the median
Full reply
20.3 s 22.6 s
1.1× better than median
Throughput
22.3 tok/s 22.3 tok/s
at the median

step-5-preview median of answering models

Error breakdown

40 failed of 158 probes, last 7 days.

FailureProbesShare of failures
Empty answer27 67.5%
Timed out10 25.0%
Not served (404)3 7.5%

Time to first token

Each probe. Lower is better.

0 ms125.0 s250.0 s375.0 s500.0 s11-10 17:0011-17 16:13 UTC
answered no answer

Throughput

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

0.0 tok/s50.0 tok/s100 tok/s150 tok/s200 tok/s11-10 17:0011-17 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
Timed outn/an/an/a
OK6.51 s16.2 s20.7 tok/s
OK7.49 s25.4 s11.2 tok/s
Timed outn/an/an/a
Timed outn/an/an/a
OK111.6 s134.4 s8.8 tok/s
OK293 ms7.40 s28.1 tok/s
Timed outn/an/an/a
Timed outn/an/an/a
OK107.8 s126.5 s10.7 tok/s
Timed outn/an/an/a
OK26.8 s57.6 s6.5 tok/s
OK8.84 s18.4 s23.9 tok/s
Empty answern/an/an/a
OK350 ms10.4 s19.8 tok/s
OK196.0 s205.9 s2.4 tok/s
OK14.2 s17.9 s28.0 tok/s
OK18.1 s22.5 s22.3 tok/s
OK4.48 s7.87 s63.5 tok/s
OK33.5 s33.6 s14.9 tok/s

Call step-5-preview

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": "stepfun-ai/step-5-preview",
    "messages": [{"role": "user", "content": "Hello"}],
    "max_tokens": 256,
    "stream": true
  }'

Other models answering

Top of the leaderboard right now.