- Writing speed, each persondecode_tps · tok/s
- 7.51 tok/sper person · median of 16
- First word afterttft_median_s · s
- 93.10 sper person · median of 16
- Slowest first wordttft_max_s · s
- 128.86 sper person · max of 16
- Machine totalaggregate_decode_tps · tok/s
- 79.19 tok/smachine total · single · sample count not recorded
- Machine total including first-word waitaggregate_e2e_tps · tok/s
- 50.49 tok/smachine total · single · sample count not recorded
- Everyone answered inwall_seconds · s
- 162.24 smachine total · single · sample count not recorded
- Peak memory, whole hostpeak_host_unavailable_gib · GiB
- 94.35 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
| Date | Shape | Each person · median | Machine total | First word · median | Result | Run |
|---|---|---|---|---|---|---|
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.
Per-person table · all measured request rows
| Person | Filled input | Cached tokens | Output | First word | Writing after first word | Total time | Checks | Run |
|---|---|---|---|---|---|---|---|---|
1 | 40,103 | unknown | 512 | 119.11 s | 11.95 tok/sone person · includes waits | 161.87 s | 3/3 Hidden facts found · completed · checks passed | |
2 | 40,107 | unknown | 512 | 105.90 s | 9.19 tok/sone person · includes waits | 161.50 s | 3/3 Hidden facts found · completed · checks passed | |
3 | 40,105 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
4 | 40,099 | unknown | 512 | 125.01 s | 13.79 tok/sone person · includes waits | 162.08 s | 3/3 Hidden facts found · completed · checks passed | |
5 | 40,103 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
6 | 40,107 | unknown | 512 | 115.21 s | 10.98 tok/sone person · includes waits | 161.75 s | 3/3 Hidden facts found · completed · checks passed | |
7 | 40,102 | unknown | 512 | 118.31 s | 11.74 tok/sone person · includes waits | 161.85 s | 3/3 Hidden facts found · completed · checks passed | |
8 | 40,103 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
9 | 40,102 | unknown | 512 | 128.86 s | 15.31 tok/sone person · includes waits | 162.24 s | 3/3 Hidden facts found · completed · checks passed | |
10 | 40,103 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
11 | 40,102 | unknown | 512 | 128.16 s | 15.01 tok/sone person · includes waits | 162.20 s | 3/3 Hidden facts found · completed · checks passed | |
12 | 40,102 | unknown | 512 | 93.10 s | 7.51 tok/sone person · includes waits | 161.13 s | 3/3 Hidden facts found · completed · checks passed | |
13 | 40,103 | unknown | 512 | 91.66 s | 7.36 tok/sone person · includes waits | 161.09 s | 3/3 Hidden facts found · completed · checks passed | |
14 | 40,102 | unknown | 512 | 58.99 s | 5.06 tok/sone person · includes waits | 160.05 s | 3/3 Hidden facts found · completed · checks passed | |
15 | 40,103 | unknown | 512 | 93.10 s | 7.51 tok/sone person · includes waits | 161.14 s | 3/3 Hidden facts found · completed · checks passed | |
16 | 40,103 | unknown | 512 | 60.74 s | 5.14 tok/sone person · includes waits | 160.11 s | 3/3 Hidden facts found · completed · checks passed |
Per-person table · all measured request rows
| Person | Filled input | Cached tokens | Output | First word | Writing after first word | Total time | Checks | Run |
|---|---|---|---|---|---|---|---|---|
1 | 40,103 | unknown | 512 | 119.11 s | 11.95 tok/sone person · includes waits | 161.87 s | 3/3 Hidden facts found · completed · checks passed | |
2 | 40,107 | unknown | 512 | 105.90 s | 9.19 tok/sone person · includes waits | 161.50 s | 3/3 Hidden facts found · completed · checks passed | |
3 | 40,105 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
4 | 40,099 | unknown | 512 | 125.01 s | 13.79 tok/sone person · includes waits | 162.08 s | 3/3 Hidden facts found · completed · checks passed | |
5 | 40,103 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
6 | 40,107 | unknown | 512 | 115.21 s | 10.98 tok/sone person · includes waits | 161.75 s | 3/3 Hidden facts found · completed · checks passed | |
7 | 40,102 | unknown | 512 | 118.31 s | 11.74 tok/sone person · includes waits | 161.85 s | 3/3 Hidden facts found · completed · checks passed | |
8 | 40,103 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
9 | 40,102 | unknown | 512 | 128.86 s | 15.31 tok/sone person · includes waits | 162.24 s | 3/3 Hidden facts found · completed · checks passed | |
10 | 40,103 | unknown | 512 | 62.32 s | 5.22 tok/sone person · includes waits | 160.18 s | 3/3 Hidden facts found · completed · checks passed | |
11 | 40,102 | unknown | 512 | 128.16 s | 15.01 tok/sone person · includes waits | 162.20 s | 3/3 Hidden facts found · completed · checks passed | |
12 | 40,102 | unknown | 512 | 93.10 s | 7.51 tok/sone person · includes waits | 161.13 s | 3/3 Hidden facts found · completed · checks passed | |
13 | 40,103 | unknown | 512 | 91.66 s | 7.36 tok/sone person · includes waits | 161.09 s | 3/3 Hidden facts found · completed · checks passed | |
14 | 40,102 | unknown | 512 | 58.99 s | 5.06 tok/sone person · includes waits | 160.05 s | 3/3 Hidden facts found · completed · checks passed | |
15 | 40,103 | unknown | 512 | 93.10 s | 7.51 tok/sone person · includes waits | 161.14 s | 3/3 Hidden facts found · completed · checks passed | |
16 | 40,103 | unknown | 512 | 60.74 s | 5.14 tok/sone person · includes waits | 160.11 s | 3/3 Hidden facts found · completed · checks passed |
Answer checks
48/48 Hidden facts found16/16 probes16/16 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
07141c2db92e47bc08777132cd1a0323faf300eab3a7d7c111bc2bf075fda0506dea9c759a0f941cf9e1cc1501216b0e107966a66c90425050e212efbd053f029758875fc55e49561165f4a44342b654c123c3e25c6811b34abb83553fb1a164- 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
- 16
- 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
| Metric | Value | Scope | Statistic | Sample count | Basis / source field | Run |
|---|---|---|---|---|---|---|
wall seconds | 162.24 s | machine total | single | not recorded | wall_seconds | |
ttft median s | 93.10 s | per person | median | 16 | ttft_median_s | |
hidden facts total | 48.00 count | machine total | single | not recorded | needle_total | |
ttft max s | 128.86 s | per person | max | 16 | ttft_max_s | |
requests ok | 16.00 count | machine total | single | not recorded | requests_ok | |
hidden facts correct | 48.00 count | machine total | single | not recorded | needle_correct | |
preemptions | 0.00 count | machine total | single | not recorded | preemptions | |
decode overlap | 16.00 count | machine total | single | not recorded | decode_overlap | |
aggregate concurrency | 16.00 count | machine total | single | not recorded | aggregate_concurrency | |
requests total | 16.00 count | machine total | single | not recorded | requests_total | |
correctness probes passed | 16.00 count | machine total | single | not recorded | correctness_probes_passed | |
aggregate decode tps | 79.19 tok/s | machine total | single | not recorded | aggregate_tps | |
aggregate e2e tps | 50.49 tok/s | machine total | single | not recorded | aggregate_end_to_end_tps | |
correctness probes total | 16.00 count | machine total | single | not recorded | correctness_probes_total | |
decode tps | 8.84 tok/s | per person | mean | 16 | per_request_decode_tps | |
forced output tokens | 512.00 tokens | machine total | single | not recorded | forced_output_tokens | |
peak host unavailable gib | 94.35 GiB | machine total | max | not recorded | peak_host_unavailable_gib | |
decode tps | 7.51 tok/s | per person | median | 16 | per_person_decode_median_tps |
Data and citation
Skilak Keep evidence, run 0bf8bb92f7a54dcab6b8fbb7, measured 2026-10-07T04:12:37.552381+00:00. https://skilak.ai/keep/resources/evidence/r/0bf8bb92f7a54dcab6b8fbb7
