vGPU-32GB · 32 GB
clef-flash
Open a report for complete measurements.
- Task
- Decision · Text → Structured data
- Workload
- 1K → 1K tokens
- Batch size
- Batch 4
- Format · Engine
- bf16 · vLLM
- Unit
- decisions/sec
- Mean throughput
- 10.9decisions/sec
- Observed GPU occupancy
- 18.5 GiB
- P95 latency
- 388 ms
- Reports combined
- 1
Launch this model
Open the decision recipe to start its API or repeat the measured batch.
Use this recipeIndividual tests
Text → Structured data · Decision · bf16 · vLLM · decision · decisions/sec
Sortable table
| Test date | Throughput | Concurrency / batch | P95 (s) | TTFT (ms) | Observed occupancy (GiB) | GPU power (W) | Decode (tok/s) | Uncached prefill (tok/s) | Parameters | View report |
|---|---|---|---|---|---|---|---|---|---|---|
| 10/7/2026, 9:44:25 AM | 10.88 decisions/sec | 4 | 0.388 | — | 18.49 | 152.78 | — | Uncached prefill unavailableView report | Parameters ↗ | caef7154 ↗ |
Recorded test parameters
Configuration 1 · Parameters
- model
- ms:Cloudflare/clef-flash
- engine
- vLLM
- quant
- bf16
- method
- native-forward-v1
- task
- decision
- dataset
- typed-invoices-v1
- context
- decision
- input_tokens
- 1024
- output_tokens
- 1024
- duration
- 1
- concurrency
- 4
- unit
- decisions/sec
- software
- —
- engine_provenance
- —
- kernel_settings
- —
- precision
- —
- resolution
- —
- performance_task
- —
- performance_scope
- —
- batch_size
- —
- dimensions
- —
- weight_quantization
- —
- max_sequence_length
- —
- model_scope
- —
- performance_dataset
- —
- workload_mode
- —
- graph_mode
- —
- model_revision
- —
- client_version
- python-native-bench-1.1.3