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 | 5 people | not recorded | not recorded | not recorded | Passed | |
2026-10-07 | 5 people | not recorded | not recorded | not recorded | Failed |
Answer checks
49/50 turns
Rolling-chat turn checks
| Person · turn | Filled input | First word | Recent · pinned · old | Result | Run |
|---|---|---|---|---|---|
1 · 1 | 20208 | 3.96 s | true · true · recalled | Passed | |
1 · 2 | 40330 | 5.72 s | true · true · recalled | Passed | |
1 · 3 | 60454 | 6.97 s | true · true · recalled | Passed | |
1 · 4 | 40508 | 9.56 s | true · true · recalled | Passed | |
1 · 5 | 60632 | 7.00 s | true · true · recalled | Passed | |
1 · 6 | 40500 | 9.58 s | true · true · recalled | Passed | |
1 · 7 | 60625 | 7.00 s | true · true · recalled | Passed | |
1 · 8 | 40501 | 13.10 s | true · true · recalled | Passed | |
1 · 9 | 60627 | 11.72 s | true · true · recalled | Passed | |
1 · 10 | 40511 | 15.42 s | true · true · recalled | Passed | |
2 · 1 | 20214 | 8.20 s | true · true · recalled | Passed | |
2 · 2 | 40334 | 11.87 s | true · true · recalled | Passed | |
2 · 3 | 60457 | 9.80 s | true · true · recalled | Passed | |
2 · 4 | 40506 | 9.61 s | true · true · recalled | Passed | |
2 · 5 | 60631 | 7.04 s | true · true · recalled | Passed | |
2 · 6 | 40505 | 9.59 s | true · true · recalled | Passed | |
2 · 7 | 60634 | 7.04 s | true · true · recalled | Passed | |
2 · 8 | 40508 | 9.57 s | true · true · recalled | Passed | |
2 · 9 | 60638 | 7.02 s | true · true · recalled | Passed | |
2 · 10 | 40510 | 9.61 s | true · true · recalled | Passed | |
3 · 1 | 20209 | 3.99 s | true · true · recalled | Passed | |
3 · 2 | 40335 | 5.73 s | true · true · recalled | Passed | |
3 · 3 | 60463 | 7.01 s | true · true · recalled | Passed | |
3 · 4 | 40503 | 9.58 s | true · true · recalled | Passed | |
3 · 5 | 60634 | 7.03 s | true · true · recalled | Passed | |
3 · 6 | 40509 | 9.59 s | true · true · recalled | Passed | |
3 · 7 | 60637 | 7.02 s | true · true · recalled | Passed | |
3 · 8 | 40503 | 9.57 s | true · true · recalled | Passed | |
3 · 9 | 60629 | 7.04 s | true · true · recalled | Passed | |
3 · 10 | 40509 | 9.60 s | true · true · recalled | Passed | |
4 · 1 | 20212 | 3.96 s | true · true · recalled | Passed | |
4 · 2 | 40337 | 5.72 s | true · true · recalled | Passed | |
4 · 3 | 60470 | 6.96 s | true · true · recalled | Passed | |
4 · 4 | 40506 | 9.55 s | true · true · recalled | Passed | |
4 · 5 | 60629 | 7.02 s | true · true · recalled | Passed | |
4 · 6 | 40501 | 9.54 s | true · true · recalled | Passed | |
4 · 7 | 60629 | 7.01 s | true · true · recalled | Passed | |
4 · 8 | 40514 | 12.78 s | true · true · recalled | Passed | |
4 · 9 | 60639 | 11.36 s | true · true · recalled | Passed | |
4 · 10 | 40514 | 15.26 s | true · true · recalled | Passed | |
5 · 1 | 20214 | 7.77 s | true · true · recalled | Passed | |
5 · 2 | 40338 | 11.44 s | true · true · incorrect | Failed | |
5 · 3 | 60464 | 9.75 s | true · true · recalled | Passed | |
5 · 4 | 40510 | 9.75 s | true · true · recalled | Passed | |
5 · 5 | 60633 | 7.04 s | true · true · recalled | Passed | |
5 · 6 | 40509 | 9.61 s | true · true · recalled | Passed | |
5 · 7 | 60639 | 7.05 s | true · true · recalled | Passed | |
5 · 8 | 40507 | 9.61 s | true · true · recalled | Passed | |
5 · 9 | 60633 | 7.05 s | true · true · recalled | Passed | |
5 · 10 | 40513 | 9.64 s | true · true · recalled | Passed |
Method
Staggered multi-turn fictional archive, exact engine-tokenized rolling gateway; independent recent, pinned and old probes each turn.
Limits
Bounded tuning pilot; incomplete window growth cannot pass the full workload grade. Lab retrieval and rolling-history evidence only. Bounded literal user-source excerpts, with verified source hashes; details outside them remain unavailable. Shared engine cache intervals include overlapping users and are not per-user hits. Gateway path, not authenticated document ingestion or customer acceptance.
What did not work
- 49/50 requests passed.
What ran
Exact identity
- Original publisher model
- Qwen3.6-35B-A3B
- Publisher revision
995ad96eacd98c81ed38be0c5b274b04031597b0- Variant · format · precision
- NVFP4 variant · safetensors · NVFP4
- Weights SHA-256
- not recorded
- Engine · version · feature
- NVIDIA NGC vLLM ·
0.29.0+5013de39.dev· plain - Hardware
- NVIDIA GB10, 128 GB unified memory
- Harness / workload
keep-rolling-context-v1 · 1- Workload suite · version
not recorded- Harness SHA-256
not recorded- Fixture SHA-256
not recorded
Workload shape
- People in this run
- 5
- Configured window (tokens)
- 262144
- Configured window (tokens)
- 262144
- Output token ceiling
- 256
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 |
|---|---|---|---|---|---|---|
roll ttft max seconds | 15.42 s | per person | max | not recorded | roll_ttft_max_seconds | |
requests ok | 49.00 count | machine total | single | not recorded | requests_ok | |
roll ttft p50 seconds | 9.60 s | per person | median | not recorded | roll_ttft_p50_seconds | |
rolled off abstained | 0.00 count | machine total | single | not recorded | rolled_off_abstained | |
rolled off invalid | 0.00 count | machine total | single | not recorded | rolled_off_invalid | |
rolled off wrong | 0.00 count | machine total | single | not recorded | rolled_off_wrong | |
normal ttft max seconds | 11.87 s | per person | max | not recorded | normal_ttft_max_seconds | |
rolled off correct | 35.00 count | machine total | single | not recorded | rolled_off_correct | |
normal ttft p50 seconds | 7.03 s | per person | median | not recorded | normal_ttft_p50_seconds | |
rolled off total | 35.00 count | machine total | single | not recorded | rolled_off_total | |
requests total | 50.00 count | machine total | single | not recorded | requests_total |
Data and citation
Skilak Keep evidence, run 90c568b5a488418c98b60943, measured 2026-10-07T19:14:45.660805+00:00. https://skilak.ai/keep/resources/evidence/r/90c568b5a488418c98b60943
