End-to-end completion throughput
39.411 tok/sObserved completion tokens, including reasoning, divided by the full measured batch duration. Includes waiting, prefill, generation and relay latency; not decode TPS.
8 samplesGLM-5.3 Flash Abliterated
abliterated · Timing-only C1; eight naturally stopping requests per input target; shared endpoint
completedPerformanceActual inference
flash-frozen-4k-c1-bc80cb97a2aaObserved outcomes
8 / 8 requests completed
Timing-only run. Answer correctness was not measured; no quality score is assigned.
8answers ungraded
First generated output, median
1,246.246 msFirst nonempty reasoning or content payload observed by the client. This is not server-internal token timing.
8 samplesFirst final content, median
1,558.713 msFirst nonempty content payload. Correctness and usefulness were not graded.
8 samplesRequest completion, median
1,858.121 msFull client-observed request duration, including shared-endpoint queueing and relay latency.
8 samplesMeasured batch duration
15.3 sClient measurement interval excluding process setup; the throughput denominator.
1 samplesCase results
Observed outcome and primary-model attribution are separate.
request-01Synthetic timing request 01Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 1,765.468 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-02Synthetic timing request 02Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 1,905.353 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-03Synthetic timing request 03Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 1,900.409 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-04Synthetic timing request 04Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 1,799.227 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-05Synthetic timing request 05Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 1,815.833 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-06Synthetic timing request 06Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 1,758.255 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-07Synthetic timing request 07Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 2,116.511 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
request-08Synthetic timing request 08Answer not graded
- Primary outcome
- unverified
- Failure stage
- answer_checks_not_run
- Duration
- 2,233.749 ms
Recorded mechanism checks
- Transport completed
- answer correctness was not graded.
Deployment and test conditions
Configured capacity is distinct from successfully tested context or concurrency.
- Model
- GLM-5.3 Flash Abliterated
- Checkpoint revision
- c8f58e6aa9117c73607d692978b22f091d80450c
- Variant
- abliterated
- Quantization
- EXL3 4 bpw weights; NVFP4 MLA KV cache
- Hardware
- Two NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation GPUs, 96 GB per GPU
- Node count
- 1
- Runtime
- vLLM
- Runtime version
- 0.26.1rc0+infernal.invocation.cu133.r19.vllm174c789.b12x12c4263
- Configured context
- 327680
- Configured request slots
- 4
- Profile
- Timing-only C1; eight naturally stopping requests per input target; shared endpoint
- Weight identity
- Not verified
- Suite version
- static-performance-evidence-1
- Started
- Sep 21, 2026, 7:14 PM UTC
- Finished
- Sep 21, 2026, 7:15 PM UTC
- Input-size target
- 4096
- Observed input-token range
- 3704–3735
- Client concurrency limit
- 1
- Observed peak request overlap
- 1
- Cache condition
- Cache state unreported
- Transport completions
- 8 / 8
- Completion-token accounting
- Reasoning and final content combined
- Stream intervals
- Not measured
- Endpoint usage
- Shared; includes relay latency. AEVAmon was paused.
- Cache condition
- Warm serving engine; no cache flush or verified prefix-cache state.
- Request limit
- One request at a time, eight requests per input target.
- Answer checks
- Not run; timing evidence only. No quality score or ranking.
- Timing observer SHA-256
- 2977430a164f9ef681d6cff4dd7e5826318f55d95d9ffe8740bce494dec505a9
- Dataset SHA-256
- 53ddfba9e794833f0ddb92f9785dc1c012c566ef6d438df0b4dd153ca074860f
- Private timing receipt SHA-256
- bc80cb97a2aaf41f078dd3e1ea23f3091e71928c478124fded3e5fc284962da2
- Observed output-token range
- 54–103
Provenance
- Suite SHA-256
- 53ddfba9e794833f0ddb92f9785dc1c012c566ef6d438df0b4dd153ca074860f
- Evaluator SHA-256
- 2977430a164f9ef681d6cff4dd7e5826318f55d95d9ffe8740bce494dec505a9
- Comparison conditions SHA-256
- Comparability not established
No additional public evidence links were attached. The run JSON contains all published fields.
Public records exclude production conversations, private memories, network endpoints and hidden model reasoning.