Report · resources

Team load

Qwen3.6-35B-A3B · NVFP4 variant · NVIDIA NGC vLLM 0.29

Measured results / Team load / Run d65163d3e20a4aac92032dc2

Measured 2026-10-07 04:19 UTC · Record date 2026-10-07

Passed. 60/60 Hidden facts found20/20 probes20/20 request checks

Writing speed, each persondecode_tps · tok/s
5.95 tok/sper person · median of 20
First word afterttft_median_s · s
114.35 sper person · median of 20
Slowest first wordttft_max_s · s
162.46 sper person · max of 20
Machine totalaggregate_decode_tps · tok/s
71.52 tok/smachine total · single · sample count not recorded
Machine total including first-word waitaggregate_e2e_tps · tok/s
50.68 tok/smachine total · single · sample count not recorded
Everyone answered inwall_seconds · s
202.05 smachine total · single · sample count not recorded
Peak memory, whole hostpeak_host_unavailable_gib · GiB
93.79 GiBmachine total · max · sample count not recorded

Runs on this setup

Earlier runs keep their own engine versions and outcomes. Workload shapes stay separate.

History and measured sweep
History of this setup, newest first
DateShapeEach person · medianMachine totalFirst word · medianResultRun
2026-10-07
20 people
5.95 tok/s
71.52 tok/s
114.35 s
Passed
2026-10-07
16 people
7.51 tok/s
79.19 tok/s
93.10 s
Passed
2026-10-07
8 people
16.80 tok/s
126.01 tok/s
54.40 s
Passed
2026-10-07
1 people
68.55 tok/s
68.55 tok/s
9.78 s
Passed

Each person in this run

First-word wait against writing speed after that word, including pauses. Earlier streams can pause while other people’s prompts are read. The table preserves filled input, cached-token counts and each person’s checks.

Each person's first word and writing speed; table below00 s785 s15171 stok/s after first wordfirst word afterPerson 1: 93.53 s · 4.80 tok/sPerson 2: 162.13 s · 12.81 tok/sPerson 3: 154.09 s · 10.72 tok/sPerson 4: 125.18 s · 6.75 tok/sPerson 5: 91.93 s · 4.73 tok/sPerson 6: 93.53 s · 4.80 tok/sPerson 7: 62.45 s · 3.75 tok/sPerson 8: 60.89 s · 3.71 tok/sPerson 9: 108.32 s · 5.55 tok/sPerson 10: 62.45 s · 3.75 tok/sPerson 11: 120.38 s · 6.36 tok/sPerson 12: 62.45 s · 3.75 tok/sPerson 13: 162.46 s · 12.91 tok/sPerson 14: 146.27 s · 9.25 tok/sPerson 15: 126.70 s · 6.88 tok/sPerson 16: 138.43 s · 8.13 tok/sPerson 17: 159.40 s · 12.02 tok/sPerson 18: 149.42 s · 9.79 tok/sPerson 19: 59.14 s · 3.67 tok/sPerson 20: 62.45 s · 3.75 tok/s
Per-person table · all measured request rows
Per-person requests for this run
PersonFilled inputCached tokensOutputFirst wordWriting after first wordTotal timeChecksRun
1
40,100
unknown
512
93.53 s
4.80 tok/sone person · includes waits
199.94 s
3/3 Hidden facts found · completed · checks passed
2
40,103
unknown
512
162.13 s
12.81 tok/sone person · includes waits
202.02 s
3/3 Hidden facts found · completed · checks passed
3
40,102
unknown
512
154.09 s
10.72 tok/sone person · includes waits
201.75 s
3/3 Hidden facts found · completed · checks passed
4
40,100
unknown
512
125.18 s
6.75 tok/sone person · includes waits
200.93 s
3/3 Hidden facts found · completed · checks passed
5
40,104
unknown
512
91.93 s
4.73 tok/sone person · includes waits
199.88 s
3/3 Hidden facts found · completed · checks passed
6
40,101
unknown
512
93.53 s
4.80 tok/sone person · includes waits
199.94 s
3/3 Hidden facts found · completed · checks passed
7
40,106
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
8
40,106
unknown
512
60.89 s
3.71 tok/sone person · includes waits
198.65 s
3/3 Hidden facts found · completed · checks passed
9
40,104
unknown
512
108.32 s
5.55 tok/sone person · includes waits
200.43 s
3/3 Hidden facts found · completed · checks passed
10
40,102
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
11
40,103
unknown
512
120.38 s
6.36 tok/sone person · includes waits
200.79 s
3/3 Hidden facts found · completed · checks passed
12
40,099
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
13
40,102
unknown
512
162.46 s
12.91 tok/sone person · includes waits
202.05 s
3/3 Hidden facts found · completed · checks passed
14
40,099
unknown
512
146.27 s
9.25 tok/sone person · includes waits
201.51 s
3/3 Hidden facts found · completed · checks passed
15
40,100
unknown
512
126.70 s
6.88 tok/sone person · includes waits
200.97 s
3/3 Hidden facts found · completed · checks passed
16
40,104
unknown
512
138.43 s
8.13 tok/sone person · includes waits
201.29 s
3/3 Hidden facts found · completed · checks passed
17
40,105
unknown
512
159.40 s
12.02 tok/sone person · includes waits
201.92 s
3/3 Hidden facts found · completed · checks passed
18
40,102
unknown
512
149.42 s
9.79 tok/sone person · includes waits
201.61 s
3/3 Hidden facts found · completed · checks passed
19
40,099
unknown
512
59.14 s
3.67 tok/sone person · includes waits
198.57 s
3/3 Hidden facts found · completed · checks passed
20
40,101
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
Per-person table · all measured request rows
Per-person requests for this run
PersonFilled inputCached tokensOutputFirst wordWriting after first wordTotal timeChecksRun
1
40,100
unknown
512
93.53 s
4.80 tok/sone person · includes waits
199.94 s
3/3 Hidden facts found · completed · checks passed
2
40,103
unknown
512
162.13 s
12.81 tok/sone person · includes waits
202.02 s
3/3 Hidden facts found · completed · checks passed
3
40,102
unknown
512
154.09 s
10.72 tok/sone person · includes waits
201.75 s
3/3 Hidden facts found · completed · checks passed
4
40,100
unknown
512
125.18 s
6.75 tok/sone person · includes waits
200.93 s
3/3 Hidden facts found · completed · checks passed
5
40,104
unknown
512
91.93 s
4.73 tok/sone person · includes waits
199.88 s
3/3 Hidden facts found · completed · checks passed
6
40,101
unknown
512
93.53 s
4.80 tok/sone person · includes waits
199.94 s
3/3 Hidden facts found · completed · checks passed
7
40,106
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
8
40,106
unknown
512
60.89 s
3.71 tok/sone person · includes waits
198.65 s
3/3 Hidden facts found · completed · checks passed
9
40,104
unknown
512
108.32 s
5.55 tok/sone person · includes waits
200.43 s
3/3 Hidden facts found · completed · checks passed
10
40,102
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
11
40,103
unknown
512
120.38 s
6.36 tok/sone person · includes waits
200.79 s
3/3 Hidden facts found · completed · checks passed
12
40,099
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed
13
40,102
unknown
512
162.46 s
12.91 tok/sone person · includes waits
202.05 s
3/3 Hidden facts found · completed · checks passed
14
40,099
unknown
512
146.27 s
9.25 tok/sone person · includes waits
201.51 s
3/3 Hidden facts found · completed · checks passed
15
40,100
unknown
512
126.70 s
6.88 tok/sone person · includes waits
200.97 s
3/3 Hidden facts found · completed · checks passed
16
40,104
unknown
512
138.43 s
8.13 tok/sone person · includes waits
201.29 s
3/3 Hidden facts found · completed · checks passed
17
40,105
unknown
512
159.40 s
12.02 tok/sone person · includes waits
201.92 s
3/3 Hidden facts found · completed · checks passed
18
40,102
unknown
512
149.42 s
9.79 tok/sone person · includes waits
201.61 s
3/3 Hidden facts found · completed · checks passed
19
40,099
unknown
512
59.14 s
3.67 tok/sone person · includes waits
198.57 s
3/3 Hidden facts found · completed · checks passed
20
40,101
unknown
512
62.45 s
3.75 tok/sone person · includes waits
198.73 s
3/3 Hidden facts found · completed · checks passed

Answer checks

60/60 Hidden facts found20/20 probes20/20 request checks

Method

team-load-v1; distinct generated 40000-token coding contexts plus fresh nonce prefixes; synchronized streams; temperature zero; thinking off; min_tokens=max_tokens=512; separate cold hidden facts probes; no retries.

Limits

Finite lab cohort; configured 262144 context differs from measured inputs. Aggregate uses first-output to final-completion window. Whole-host memory sampled during speed only; missing counters remain unassessed. Broader coding quality and delivery acceptance unassessed.

What ran

Exact identity

Original publisher model
Qwen3.6-35B-A3B
Publisher revision
995ad96eacd98c81ed38be0c5b274b04031597b0
Variant · format · precision
NVFP4 variant · safetensors · nvfp4
Weights SHA-256
07141c2db92e47bc08777132cd1a0323faf300eab3a7d7c111bc2bf075fda050
6dea9c759a0f941cf9e1cc1501216b0e107966a66c90425050e212efbd053f02
9758875fc55e49561165f4a44342b654c123c3e25c6811b34abb83553fb1a164
Engine · version · feature
NVIDIA NGC vLLM · 0.29.0+5013de39.dev · plain
Hardware
NVIDIA GB10, 128 GB unified memory
Harness / workload
keep-team-load-v1 · 1
Workload suite · version
not recorded
Harness SHA-256
not recorded
Fixture SHA-256
e2097959b1ca0142f2b93edc3ec411e474c0a6622800dd247af7a71bd9c39e83

Workload shape

People in this run
20
Configured window (tokens)
262144
Configured window (tokens)
262144
KV cache precision
fp8
Output token ceiling
512
Feature
plain
Thinking
off

Full reproduction pack available to customers and evaluators on request

All recorded numbers

Metric definitions, scope and statistics
All aggregate numeric result fields in this run
MetricValueScopeStatisticSample countBasis / source fieldRun
wall seconds
202.05 s
machine total
single
not recorded
wall_seconds
ttft median s
114.35 s
per person
median
20
ttft_median_s
hidden facts total
60.00 count
machine total
single
not recorded
needle_total
ttft max s
162.46 s
per person
max
20
ttft_max_s
requests ok
20.00 count
machine total
single
not recorded
requests_ok
hidden facts correct
60.00 count
machine total
single
not recorded
needle_correct
preemptions
0.00 count
machine total
single
not recorded
preemptions
decode overlap
20.00 count
machine total
single
not recorded
decode_overlap
aggregate concurrency
20.00 count
machine total
single
not recorded
aggregate_concurrency
requests total
20.00 count
machine total
single
not recorded
requests_total
correctness probes passed
20.00 count
machine total
single
not recorded
correctness_probes_passed
aggregate decode tps
71.52 tok/s
machine total
single
not recorded
aggregate_tps
aggregate e2e tps
50.68 tok/s
machine total
single
not recorded
aggregate_end_to_end_tps
correctness probes total
20.00 count
machine total
single
not recorded
correctness_probes_total
decode tps
6.89 tok/s
per person
mean
20
per_request_decode_tps
forced output tokens
512.00 tokens
machine total
single
not recorded
forced_output_tokens
peak host unavailable gib
93.79 GiB
machine total
max
not recorded
peak_host_unavailable_gib
decode tps
5.95 tok/s
per person
median
20
per_person_decode_median_tps

Data and citation

Skilak Keep evidence, run d65163d3e20a4aac92032dc2, measured 2026-10-07T04:19:07.006640+00:00.
https://skilak.ai/keep/resources/evidence/r/d65163d3e20a4aac92032dc2

Ready to scope it?

Start with one workload.

Tell us who uses it, what it can read, and what must stay inside. We will map the build.