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.

deepseek

deepseek-v4.1-flash

Unstable

Rank 15 of 15 Last probed (22 min ago)

deepseek-ai/deepseek-v4.1-flash answered 42.5% of probes on the NVIDIA NIM free API in the last 14 days, with a median time to first token of 199.4 s and 1.8 tok/s. It ranks 15 of 15 tracked models.

42.5% Uptime 138/325 probes
28/ 100 Score 0 round wins
199.4 s First token p95 319.0 s
209.5 s Full reply median
8.90 s Best reply fastest in range
1.8tok/s Throughput median

Model vs. median

Against the median of models that answered, last 14 days.

Uptime
42.5% 79.4%
37 pts below median
First token
199.4 s 13.0 s
15× worse than median
Full reply
209.5 s 22.6 s
9.3× worse than median
Throughput
1.8 tok/s 21.3 tok/s
12× worse than median

deepseek-v4.1-flash median of answering models

Error breakdown

187 failed of 325 probes, last 14 days.

FailureProbesShare of failures
Timed out138 73.8%
Network error24 12.8%
Empty answer13 7.0%
Not served (404)12 6.4%

Time to first token

Each probe. Lower is better.

0 ms125.0 s250.0 s375.0 s500.0 s11-03 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/s12.5 tok/s25.0 tok/s37.5 tok/s50.0 tok/s11-03 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
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Not served (404)n/an/an/a
Timed outn/an/an/a
Timed outn/an/an/a
Timed outn/an/an/a
Timed outn/an/an/a
OK289.3 s291.6 s1.3 tok/s
Timed outn/an/an/a
Timed outn/an/an/a
Empty answern/an/an/a

Call deepseek-v4.1-flash

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": "deepseek-ai/deepseek-v4.1-flash",
    "messages": [{"role": "user", "content": "Hello"}],
    "max_tokens": 256,
    "stream": true
  }'

Other models answering

Top of the leaderboard right now.