32 GB · 53 reports

vGPU-32GB · 32 GB

Qualified results first; original reports and full tables stay below.

All tasks · Qualified results

Speeds compare only within one exact test. Recipe availability is checked on each report against the published recipe executor.

embeddinggemma-2

google · 2 tested configurations

bf16 · Text → Vectors · embedding-mteb-tasks-v1 · Test duration 9 s

Throughput · Batch 1: 251vectors/sec
Best measured · Batch 16: 2,710vectors/sec
Observed GPU occupancy: 1GiB
P95 latency: 4 ms

Use this recipe · Batch 16 ↗Evidence report and run availability ↗Same test on other GPUs

Test details · 3 batch sizes

Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.

Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.

Engine: sentence-transformers · PyTorch 2.13.0+cu130
Protocol: native-forward-v1
Dataset: embedding-mteb-tasks-v1
Workload: embedding

Batch size	Mean throughput (vectors/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	251.11	4 ms	—	1.05	99.4	1	f489c748 ↗
B4	924.83	4 ms	—	1.05	111.5	1	f489c748 ↗
B16	2,710.19	6 ms	—	1.05	114.9	1	f489c748 ↗

1 other test configurations on this GPU

bf16 · Text → Vectors · embedding-sanity-v1 · Test duration 12 s

Throughput · Batch 1: 38.8vectors/sec
Best measured · Batch 64: 1,766vectors/sec
Observed GPU occupancy: 0.8GiB
P95 latency: 27 ms

Use this recipe · Batch 64 ↗Evidence report and run availability ↗Same test on other GPUs

Test details · 4 batch sizes

Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.

Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.

Engine: sentence-transformers · PyTorch 2.13.0+cu130
Protocol: native-forward-v1
Dataset: embedding-sanity-v1
Workload: embedding

Batch size	Mean throughput (vectors/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	38.8	27 ms	—	0.82	48.8	1	dc5960cd ↗
B4	141.65	29 ms	—	0.82	52.4	1	dc5960cd ↗
B16	553.54	30 ms	—	0.87	56.4	1	dc5960cd ↗
B64	1,766.47	38 ms	—	1.03	65.3	1	dc5960cd ↗

Kokoro-82M
hexgrad · 3 tested configurations
- fp32 · Text → Audio · kokoro-english-v1 · Test duration 6 s
  Throughput · Batch 1
  76.7audio-sec/sec
  Best measured · Batch 2
  79.8audio-sec/sec
  Observed GPU occupancy
  1GiB
  P95 latency
  45 ms
  Evidence report and run availability ↗Same test on other GPUs
  Test details · 2 batch sizes
  Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
  Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
  Engine
  kokoro · PyTorch 2.13.0+cu130
  Protocol
  audio-native-v1
  Dataset
  kokoro-english-v1
  Workload
  tts
  Batch size Mean throughput (audio-sec/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
  B1 76.72 45 ms — 1 83.5 1 d90a5c29 ↗
  B2 79.84 100 ms — 1.03 100.8 1 d90a5c29 ↗
- 2 other test configurations on this GPU
  - fp32 · Text → Audio · kokoro-english-v1 · Test duration 6 s
    Throughput · Batch 1
    40.4audio-sec/sec
    Best measured · Batch 4
    41.8audio-sec/sec
    Observed GPU occupancy
    1GiB
    P95 latency
    104 ms
    Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    kokoro · PyTorch 2.13.0+cu130
    Protocol
    audio-native-v1
    Dataset
    kokoro-english-v1
    Workload
    tts
    Batch size Mean throughput (audio-sec/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 40.36 104 ms — 0.99 60.5 1 c0488108 ↗
    B4 41.8 426 ms — 1.03 65.9 1 c0488108 ↗
  - fp32 · Text → Audio · kokoro-english-v1 · Test duration 6 s
    Throughput · Batch 1
    43.4audio-sec/sec
    Best measured · Batch 4
    44.9audio-sec/sec
    Observed GPU occupancy
    1GiB
    P95 latency
    101 ms
    Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    kokoro
    Protocol
    audio-native-v1
    Dataset
    kokoro-english-v1
    Workload
    tts
    Batch size Mean throughput (audio-sec/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 43.43 101 ms — 0.99 64.1 1 d8e4b3ce ↗
    B4 44.88 396 ms — 1.03 69.1 1 d8e4b3ce ↗
faster-whisper-tiny
Systran · 1 tested configuration
- fp16 · Audio → Text · asr_prompts · Test duration 1 s
  Throughput · Batch 1
  137audio-sec/sec
  Best measured · Batch 4
  205audio-sec/sec
  Observed GPU occupancy
  0.4GiB
  P95 latency
  Not recorded
  Evidence report and run availability ↗Same test on other GPUs
  Test details · 2 batch sizes
  Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
  Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
  Engine
  faster-whisper · PyTorch 2.13.0+cu130
  Protocol
  audio-native-v1
  Dataset
  asr_prompts
  Workload
  asr
  Batch size Mean throughput (audio-sec/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
  B1 137.01 — — 0.38 54.3 1 4b83f25d ↗
  B4 205.27 — — 0.38 66.3 1 4b83f25d ↗
whisper-tiny
openai · 2 tested configurations
- fp16 · Audio → Text · asr_prompts
  Throughput · Batch 1
  73.7audio-sec/sec
  Observed GPU occupancy
  0.4GiB
  P95 latency
  Not recorded
  Evidence report and run availability ↗Same test on other GPUs
  Test details · 2 batch sizes
  Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
  Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
  Engine
  faster-whisper · PyTorch 2.13.0+cu130
  Protocol
  audio-native-v1
  Dataset
  asr_prompts
  Workload
  asr
  Batch size Mean throughput (audio-sec/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
  B1 73.68 — — 0.38 46.8 1 7d182de2 ↗
  B4 68.2 — — 0.38 49.2 1 7d182de2 ↗
- 1 other test configurations on this GPU
  - fp16 · Audio → Text · asr_prompts · Test duration 4 s
    Throughput · Batch 1
    122audio-sec/sec
    Best measured · Batch 4
    168audio-sec/sec
    Observed GPU occupancy
    0.4GiB
    P95 latency
    Not recorded
    Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    faster-whisper
    Protocol
    audio-native-v1
    Dataset
    asr_prompts
    Workload
    asr
    Batch size Mean throughput (audio-sec/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 121.9 — — 0.42 73.6 1 7aaa1f0d ↗
    B4 168.25 — — 0.42 77.4 1 7aaa1f0d ↗
clef-flash
ms:Cloudflare · 4 tested configurations
- bf16 · Text → Structured data · typed-invoices-v1 · Test duration 6 s
  Throughput · Batch 1
  10.6decisions/sec
  Best measured · Batch 4
  11decisions/sec
  Observed GPU occupancy
  18.2GiB
  P95 latency
  99 ms
  Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
  Test details · 2 batch sizes
  Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
  Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
  Engine
  clef-native · PyTorch 2.13.0+cu130
  Protocol
  native-forward-v1
  Dataset
  typed-invoices-v1
  Workload
  decision
  Batch size Mean throughput (decisions/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
  B1 10.56 99 ms — 18.17 206.1 2 63f44036 ↗
  B4 11.04 402 ms — 18.55 244.4 2 63f44036 ↗
- 3 other test configurations on this GPU
  - bf16 · Text → Structured data · typed-invoices-v1 · Test duration 6 s
    Throughput · Batch 1
    10.6decisions/sec
    Best measured · Batch 4
    11.1decisions/sec
    Observed GPU occupancy
    18.2GiB
    P95 latency
    103 ms
    Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    clef-native
    Protocol
    native-forward-v1
    Dataset
    typed-invoices-v1
    Workload
    decision
    Batch size Mean throughput (decisions/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 10.56 103 ms — 18.17 196.8 1 4eb73a4d ↗
    B4 11.1 390 ms — 18.56 235.9 1 4eb73a4d ↗
  - bf16 · Text → Structured data · typed-invoices-v1 · Test duration 1 s
    Throughput · Batch 1
    10.3decisions/sec
    Best measured · Batch 4
    10.9decisions/sec
    Observed GPU occupancy
    18.2GiB
    P95 latency
    104 ms
    Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    clef-native
    Protocol
    native-forward-v1
    Dataset
    typed-invoices-v1
    Workload
    decision
    Batch size Mean throughput (decisions/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 10.26 104 ms — 18.17 67.9 1 1d1661ad ↗
    B4 10.89 389 ms — 18.49 156 1 1d1661ad ↗
  - bf16 · 1K → 1K tokens
    Throughput · Batch 1
    10.2decisions/sec
    Best measured · Batch 4
    10.9decisions/sec
    Observed GPU occupancy
    18.2GiB
    P95 latency
    104 ms
    Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    vLLM
    Protocol
    native-forward-v1
    Dataset
    typed-invoices-v1
    Workload
    decision
    Batch size Mean throughput (decisions/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 10.19 104 ms — 18.17 69.8 1 caef7154 ↗
    B4 10.88 388 ms — 18.49 152.8 1 caef7154 ↗
embeddinggemma-2
ms:google · 3 tested configurations
- bf16 · Text → Vectors · embedding-sanity-v1 · Test duration 6 s
  Throughput · Batch 1
  38.3vectors/sec
  Best measured · Batch 4
  139vectors/sec
  Observed GPU occupancy
  0.8GiB
  P95 latency
  31 ms
  Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
  Test details · 2 batch sizes
  Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
  Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
  Engine
  sentence-transformers · PyTorch 2.13.0+cu130
  Protocol
  native-forward-v1
  Dataset
  embedding-sanity-v1
  Workload
  embedding
  Batch size Mean throughput (vectors/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
  B1 38.33 31 ms — 0.82 50.5 4 2c17d254 ↗
  B4 138.64 33 ms — 0.82 54.2 4 2c17d254 ↗
- 2 other test configurations on this GPU
  - bf16 · Text → Vectors · embedding-sanity-v1 · Test duration 6 s
    Throughput · Batch 1
    38.3vectors/sec
    Best measured · Batch 4
    138vectors/sec
    Observed GPU occupancy
    0.8GiB
    P95 latency
    32 ms
    Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    sentence-transformers
    Protocol
    native-forward-v1
    Dataset
    embedding-sanity-v1
    Workload
    embedding
    Batch size Mean throughput (vectors/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 38.27 32 ms — 0.82 48.6 1 47f71b64 ↗
    B4 137.82 35 ms — 0.82 51.9 1 47f71b64 ↗
  - bf16 · Text → Vectors · embedding-sanity-v1
    Throughput · Batch 1
    39.4vectors/sec
    Best measured · Batch 4
    120vectors/sec
    Observed GPU occupancy
    Not recorded
    P95 latency
    26 ms
    Use this recipe · Batch 4 ↗Evidence report and run availability ↗Same test on other GPUs
    Test details · 2 batch sizes
    Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.
    Reports combined counts reports, not inference requests. Request successes and failures are listed in each report.
    Engine
    sentence-transformers
    Protocol
    native-forward-v1
    Dataset
    embedding-sanity-v1
    Workload
    embedding
    Batch size Mean throughput (vectors/sec) P95 TTFT Observed occupancy (GiB) GPU power (W) Reports combined View report
    B1 39.4 26 ms — — — 1 2c97c869 ↗
    B4 119.85 44 ms — — — 1 2c97c869 ↗

Batch size	Mean throughput (audio-sec/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	76.72	45 ms	—	1	83.5	1	d90a5c29 ↗
B2	79.84	100 ms	—	1.03	100.8	1	d90a5c29 ↗

Batch size	Mean throughput (audio-sec/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	40.36	104 ms	—	0.99	60.5	1	c0488108 ↗
B4	41.8	426 ms	—	1.03	65.9	1	c0488108 ↗

Batch size	Mean throughput (audio-sec/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	43.43	101 ms	—	0.99	64.1	1	d8e4b3ce ↗
B4	44.88	396 ms	—	1.03	69.1	1	d8e4b3ce ↗

Batch size	Mean throughput (audio-sec/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	137.01	—	—	0.38	54.3	1	4b83f25d ↗
B4	205.27	—	—	0.38	66.3	1	4b83f25d ↗

Batch size	Mean throughput (audio-sec/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	73.68	—	—	0.38	46.8	1	7d182de2 ↗
B4	68.2	—	—	0.38	49.2	1	7d182de2 ↗

Batch size	Mean throughput (audio-sec/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	121.9	—	—	0.42	73.6	1	7aaa1f0d ↗
B4	168.25	—	—	0.42	77.4	1	7aaa1f0d ↗

Batch size	Mean throughput (decisions/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	10.56	99 ms	—	18.17	206.1	2	63f44036 ↗
B4	11.04	402 ms	—	18.55	244.4	2	63f44036 ↗

Batch size	Mean throughput (decisions/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	10.56	103 ms	—	18.17	196.8	1	4eb73a4d ↗
B4	11.1	390 ms	—	18.56	235.9	1	4eb73a4d ↗

Batch size	Mean throughput (decisions/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	10.26	104 ms	—	18.17	67.9	1	1d1661ad ↗
B4	10.89	389 ms	—	18.49	156	1	1d1661ad ↗

Batch size	Mean throughput (decisions/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	10.19	104 ms	—	18.17	69.8	1	caef7154 ↗
B4	10.88	388 ms	—	18.49	152.8	1	caef7154 ↗

Batch size	Mean throughput (vectors/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	38.33	31 ms	—	0.82	50.5	4	2c17d254 ↗
B4	138.64	33 ms	—	0.82	54.2	4	2c17d254 ↗

Batch size	Mean throughput (vectors/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	38.27	32 ms	—	0.82	48.6	1	47f71b64 ↗
B4	137.82	35 ms	—	0.82	51.9	1	47f71b64 ↗

Batch size	Mean throughput (vectors/sec)	P95	TTFT	Observed occupancy (GiB)	GPU power (W)	Reports combined	View report
B1	39.4	26 ms	—	—	—	1	2c97c869 ↗
B4	119.85	44 ms	—	—	—	1	2c97c869 ↗

Occupancy includes engine buffers and reserved KV cache. It is not the minimum VRAM required by the model.

Sortable tables by exact test condition · 33 rows

C: concurrency · B: batch · —: not measured

Decision · 1K → 1K tokens · Batch 1 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	vLLM	decision1,024 → 1,024 tokensnative-forward-v1	B1	10.19	10.19	—	0.104	18.17	69.8	10.19	0.15	1 ↗

Decision · 1K → 1K tokens · Batch 4 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	vLLM	decision1,024 → 1,024 tokensnative-forward-v1	B4	10.88	10.88	—	0.388	18.49	152.8	10.88	0.07	1 ↗

Decision · Text → Structured data · typed-invoices-v1 · Test duration 1 s · Batch 1 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	clef-native	decisionnative-forward-v1	B1	10.26	10.26	—	0.104	18.17	67.9	10.26	0.15	1 ↗

Decision · Text → Structured data · typed-invoices-v1 · Test duration 1 s · Batch 4 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	clef-native	decisionnative-forward-v1	B4	10.89	10.89	—	0.389	18.49	156	10.89	0.07	1 ↗

Decision · Text → Structured data · typed-invoices-v1 · Test duration 6 s · Batch 1 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	clef-native	decisionnative-forward-v1	B1	10.56	10.65	—	0.099	18.17	206.1	10.48	0.05	2 ↗

Decision · Text → Structured data · typed-invoices-v1 · Test duration 6 s · Batch 1 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	clef-native	decisionnative-forward-v1	B1	10.56	10.56	—	0.103	18.17	196.8	10.56	0.05	1 ↗

Decision · Text → Structured data · typed-invoices-v1 · Test duration 6 s · Batch 4 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	clef-native	decisionnative-forward-v1	B4	11.04	11.07	—	0.402	18.55	244.4	11.02	0.05	2 ↗

Decision · Text → Structured data · typed-invoices-v1 · Test duration 6 s · Batch 4 · decisions/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:Cloudflare/clef-flashScaling curves ↗	bf16	clef-native	decisionnative-forward-v1	B4	11.1	11.1	—	0.39	18.56	235.9	11.1	0.05	1 ↗

Embedding · Text → Vectors · embedding-mteb-tasks-v1 · Test duration 9 s · Batch 1 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B1	251.11	251.11	—	0.004	1.05	99.4	251.11	2.53	1 ↗

Embedding · Text → Vectors · embedding-mteb-tasks-v1 · Test duration 9 s · Batch 4 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B4	924.83	924.83	—	0.004	1.05	111.5	924.83	8.29	1 ↗

Embedding · Text → Vectors · embedding-mteb-tasks-v1 · Test duration 9 s · Batch 16 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B16	2,710.19	2,710.19	—	0.006	1.05	114.9	2,710.19	23.58	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Batch 1 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B1	39.4	39.4	—	0.026	—	—	39.4	—	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Batch 4 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B4	119.85	119.85	—	0.044	—	—	119.85	—	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 6 s · Batch 1 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B1	38.33	40.57	—	0.031	0.82	50.5	36.18	0.76	4 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 6 s · Batch 1 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B1	38.27	38.27	—	0.032	0.82	48.6	38.27	0.79	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 6 s · Batch 4 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B4	138.64	140.72	—	0.033	0.82	54.2	134.59	2.56	4 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 6 s · Batch 4 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
ms:google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B4	137.82	137.82	—	0.035	0.82	51.9	137.82	2.66	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 12 s · Batch 1 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B1	38.8	38.8	—	0.027	0.82	48.8	38.8	0.79	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 12 s · Batch 4 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B4	141.65	141.65	—	0.029	0.82	52.4	141.65	2.7	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 12 s · Batch 16 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B16	553.54	553.54	—	0.03	0.87	56.4	553.54	9.82	1 ↗

Embedding · Text → Vectors · embedding-sanity-v1 · Test duration 12 s · Batch 64 · vectors/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
google/embeddinggemma-2Scaling curves ↗	bf16	sentence-transformers	embeddingnative-forward-v1	B64	1,766.47	1,766.47	—	0.038	1.03	65.3	1,766.47	27.07	1 ↗

Speech recognition · Audio → Text · asr_prompts · Batch 1 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
openai/whisper-tinyScaling curves ↗	fp16	faster-whisper	asraudio-native-v1	B1	73.68	73.68	—	—	0.38	46.8	73.68	1.57	1 ↗

Speech recognition · Audio → Text · asr_prompts · Batch 4 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
openai/whisper-tinyScaling curves ↗	fp16	faster-whisper	asraudio-native-v1	B4	68.2	68.2	—	—	0.38	49.2	68.2	1.39	1 ↗

Speech recognition · Audio → Text · asr_prompts · Test duration 1 s · Batch 1 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
Systran/faster-whisper-tinyScaling curves ↗	fp16	faster-whisper	asraudio-native-v1	B1	137.01	137.01	—	—	0.38	54.3	137.01	2.52	1 ↗

Speech recognition · Audio → Text · asr_prompts · Test duration 1 s · Batch 4 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
Systran/faster-whisper-tinyScaling curves ↗	fp16	faster-whisper	asraudio-native-v1	B4	205.27	205.27	—	—	0.38	66.3	205.27	3.1	1 ↗

Speech recognition · Audio → Text · asr_prompts · Test duration 4 s · Batch 1 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
openai/whisper-tinyScaling curves ↗	fp16	faster-whisper	asraudio-native-v1	B1	121.9	121.9	—	—	0.42	73.6	121.9	1.66	1 ↗

Speech recognition · Audio → Text · asr_prompts · Test duration 4 s · Batch 4 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
openai/whisper-tinyScaling curves ↗	fp16	faster-whisper	asraudio-native-v1	B4	168.25	168.25	—	—	0.42	77.4	168.25	2.17	1 ↗

Speech synthesis · Text → Audio · kokoro-english-v1 · Test duration 6 s · Batch 1 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
hexgrad/Kokoro-82MScaling curves ↗	fp32	kokoro	ttsaudio-native-v1	B1	76.72	76.72	—	0.045	1	83.5	76.72	0.92	1 ↗

Speech synthesis · Text → Audio · kokoro-english-v1 · Test duration 6 s · Batch 1 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
hexgrad/Kokoro-82MScaling curves ↗	fp32	kokoro	ttsaudio-native-v1	B1	40.36	40.36	—	0.104	0.99	60.5	40.36	0.67	1 ↗

Speech synthesis · Text → Audio · kokoro-english-v1 · Test duration 6 s · Batch 1 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
hexgrad/Kokoro-82MScaling curves ↗	fp32	kokoro	ttsaudio-native-v1	B1	43.43	43.43	—	0.101	0.99	64.1	43.43	0.68	1 ↗

Speech synthesis · Text → Audio · kokoro-english-v1 · Test duration 6 s · Batch 2 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
hexgrad/Kokoro-82MScaling curves ↗	fp32	kokoro	ttsaudio-native-v1	B2	79.84	79.84	—	0.1	1.03	100.8	79.84	0.79	1 ↗

Speech synthesis · Text → Audio · kokoro-english-v1 · Test duration 6 s · Batch 4 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
hexgrad/Kokoro-82MScaling curves ↗	fp32	kokoro	ttsaudio-native-v1	B4	41.8	41.8	—	0.426	1.03	65.9	41.8	0.63	1 ↗

Speech synthesis · Text → Audio · kokoro-english-v1 · Test duration 6 s · Batch 4 · audio-sec/sec · 1 rows
Model	Format	Engine	Workload	Concurrency / batch	Mean throughput	Best measured	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	GPU power (W)	Lowest measured	Throughput / W	Reports combined
hexgrad/Kokoro-82MScaling curves ↗	fp32	kokoro	ttsaudio-native-v1	B4	44.88	44.88	—	0.396	1.03	69.1	44.88	0.65	1 ↗

Scaling curves and individual runs

One model and test protocol per curve; different workloads are never joined into one line.

Benchmark configuration

google/embeddinggemma-2 · bf16 · sentence-transformers · embedding · native-forward-v1

Mean throughput vectors/sec

Measured points only. Select a point to open its source report.

Mean run P95 latency s

Measured points only. Select a point to open its source report.

Mean GPU memory used GB

Measured points only. Select a point to open its source report.

Each point is the mean of recorded runs at that level. A source report may differ from the mean; all runs are listed below.

Runs behind the curve · google/embeddinggemma-2 · vectors/sec · 3 rows
Batch size	Throughput	Decode (tok/s)	Uncached prefill (tok/s)	GPU power (W)	TTFT (ms)	P95 (s)	Observed occupancy (GiB)	Test date	View report
1	251.11 vectors/sec	—	Uncached prefill unavailableView report	99.39	—	0.004	1.05	10/8/2026, 9:01:22 AM	f489c748 ↗
4	924.83 vectors/sec	—	Uncached prefill unavailableView report	111.54	—	0.004	1.05	10/8/2026, 9:01:22 AM	f489c748 ↗
16	2,710.19 vectors/sec	—	Uncached prefill unavailableView report	114.93	—	0.006	1.05	10/8/2026, 9:01:22 AM	f489c748 ↗

Quality by dataset

Scores from different datasets and sample scopes are not directly comparable.

Quality by dataset · 8 rows
Model	Dataset	Format	Engine	Average score	Sample size	Runs
Qwen/Qwen3.5-9B	aime2026-test-zero-shot-strict-v1-n30-mt4096-t0	FP8	openai-compatible	23.33%	30 / 30 · Full dataset	1 runs 23.33% · b597a5d2 ↗
Qwen/Qwen3.5-9B	mmlu-test-zero-shot-strict-v1-n14042-mt256-t0	FP8	openai-compatible	77.85%	14042 / 14042 · Full dataset	1 runs 77.85% · cb8d4e0e ↗
Qwen/Qwen3.5-9B	gsm8k-test-zero-shot-strict-v1-n1319-mt1024-t0	FP8	openai-compatible	81.05%	1319 / 1319 · Full dataset	1 runs 81.05% · b9209027 ↗
Qwen3-0.6B	mmlu-test-zero-shot-strict-v1-n14042-mt256-t0	Native	openai-compatible	18%	14042 / 14042 · Full dataset	1 runs 18% · 85b6da31 ↗
Qwen3-0.6B	gsm8k-test-zero-shot-strict-v1-n1319-mt1024-t0	Native	openai-compatible	25.32%	1319 / 1319 · Full dataset	1 runs 25.32% · 96e3edea ↗
Qwen3-0.6B	aime2026-test-zero-shot-strict-v1-n30-mt4096-t0	Native	openai-compatible	0%	30 / 30 · Full dataset	1 runs 0% · 156bea9e ↗
Qwen3-0.6B	mmlu-test-zero-shot-strict-v1-n32-mt256-t0	Native	openai-compatible	15.63%	32 / 14042 · Subset only	1 runs 15.63% · e3289156 ↗
Qwen3-0.6B	gsm8k-test-zero-shot-strict-v1-n32-mt1024-t0	Native	openai-compatible	12.5%	32 / 1319 · Subset only	1 runs 12.5% · 184c7d7f ↗

Measurements without a comparable protocol · 7 rows

These measured rates remain visible, but are excluded from comparison curves because the protocol is unknown.

Measurements without a comparable protocol · 7 rows
Model	Format	Engine	Workload	Concurrency	Throughput	GPU power (W)
Qwen3-0.6B ↗	Native	vLLM	512	1	310.63 tok/s	163
Qwen3.5-0.8B ↗	BF16	vLLM	512	1	284.63 tok/s	171.69
Qwen3.5-0.8B ↗	BF16	vLLM	512	4	996.7 tok/s	153.72
Qwen3.5-2B ↗	BF16	vLLM	512	1	155.66 tok/s	198.97
Qwen3.5-2B ↗	BF16	vLLM	512	4	521.83 tok/s	175.87
Qwen3.5-4B ↗	BF16	vLLM	512	1	71.33 tok/s	225.14
Qwen3.5-4B ↗	BF16	vLLM	512	4	243.9 tok/s	200.14

Original reports

Every report for this task, including quality, diagnostic and legacy runs that do not enter comparisons.

Original reports · 53 rows
Model	Test task	Format	Engine	Workload	Protocol	Quality score	Test date
clef-flash ↗	Decision	bf16	clef-native	decision	native-forward-v1	—	10/7/2026, 12:59:49 PM
clef-flash ↗	Decision	bf16	clef-native	decision	native-forward-v1	—	10/7/2026, 12:51:07 PM
clef-flash ↗	Decision	bf16	clef-native	decision	native-forward-v1	—	10/7/2026, 11:11:06 AM
clef-flash ↗	Decision	bf16	clef-native	decision	native-forward-v1	—	10/7/2026, 10:22:12 AM
clef-flash ↗	Decision	bf16	clef-native	decision	native-forward-v1	—	10/7/2026, 9:55:46 AM
clef-flash ↗	Decision	bf16	vLLM	decision	native-forward-v1	—	10/7/2026, 9:44:25 AM
DeepSeek-R1-Distill-Qwen-1.5B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/8/2026, 6:09:34 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/8/2026, 9:01:22 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/8/2026, 3:35:39 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/7/2026, 12:50:28 PM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/7/2026, 11:13:47 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/7/2026, 11:10:25 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/7/2026, 10:43:37 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/7/2026, 10:21:19 AM
embeddinggemma-2 ↗	Embedding	bf16	sentence-transformers	embedding	native-forward-v1	—	10/7/2026, 9:34:33 AM
faster-whisper-tiny ↗	Speech recognition	fp16	faster-whisper	asr	audio-native-v1	—	10/8/2026, 3:28:14 AM
Kokoro-82M ↗	Speech synthesis	fp32	kokoro	tts	audio-native-v1	—	10/8/2026, 3:30:29 AM
Kokoro-82M ↗	Speech synthesis	fp32	kokoro	tts	audio-native-v1	—	10/7/2026, 12:54:04 PM
Kokoro-82M ↗	Speech synthesis	fp32	kokoro	tts	audio-native-v1	—	10/7/2026, 10:23:16 AM
Kokoro-82M ↗	Speech synthesis	fp32	vLLM	tts	audio-native-v1	—	10/7/2026, 9:40:16 AM
Qwen3-0.6B ↗	Text generation	Native	openai-compatible	mmlu-test-zero-shot-strict-v1-n14042-mt256-t0	quality-v1	18%	10/8/2026, 7:47:16 AM
Qwen3-0.6B ↗	Text generation	Native	openai-compatible	gsm8k-test-zero-shot-strict-v1-n1319-mt1024-t0	quality-v1	25.32%	10/8/2026, 7:39:22 AM
Qwen3-0.6B ↗	Text generation	Native	openai-compatible	aime2026-test-zero-shot-strict-v1-n30-mt4096-t0	quality-v1	0%	10/8/2026, 7:35:26 AM
Qwen3-0.6B ↗	Text generation	Native	openai-compatible	mmlu-test-zero-shot-strict-v1-n32-mt256-t0	quality-v1	15.63%	10/8/2026, 7:30:41 AM
Qwen3-0.6B ↗	Text generation	Native	openai-compatible	gsm8k-test-zero-shot-strict-v1-n32-mt1024-t0	quality-v1	12.5%	10/8/2026, 7:30:36 AM
Qwen3-0.6B ↗	Text generation	BF16	vLLM	2K	wbench-http	—	10/8/2026, 5:19:21 AM
Qwen3-0.6B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/8/2026, 5:05:49 AM
Qwen3-0.6B ↗	Text generation	Native	vLLM	512	Method unknown	—	10/6/2026, 11:42:32 PM
Qwen3-0.6B-FP8 ↗	Text generation	FP8	vLLM	2K	wbench-http	—	10/8/2026, 4:20:58 AM
Qwen3-0.6B-FP8 ↗	Text generation	FP8	vLLM	512	wbench-http	—	10/8/2026, 4:08:15 AM
Qwen3-0.6B-GPTQ ↗	Text generation	GPTQ	vLLM	2K	wbench-http	—	10/8/2026, 4:51:08 AM
Qwen3-0.6B-GPTQ ↗	Text generation	GPTQ	vLLM	512	wbench-http	—	10/8/2026, 4:35:36 AM
Qwen3-0.6B-Q8 ↗	Text generation	Q8_0	llama.cpp	512	wbench-http	—	10/7/2026, 8:12:16 AM
Qwen3.5-0.8B ↗	Text generation	BF16	vLLM	2K	wbench-http	—	10/8/2026, 5:52:20 AM
Qwen3.5-0.8B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/8/2026, 5:34:20 AM
Qwen3.5-0.8B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/7/2026, 2:06:36 AM
Qwen3.5-0.8B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/7/2026, 2:04:58 AM
Qwen3.5-0.8B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/7/2026, 1:58:36 AM
Qwen3.5-0.8B ↗	Text generation	BF16	vLLM	512	Method unknown	—	10/7/2026, 1:08:27 AM
Qwen3.5-2B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/7/2026, 3:00:54 AM
Qwen3.5-2B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/7/2026, 2:16:29 AM
Qwen3.5-2B ↗	Text generation	BF16	vLLM	512	Method unknown	—	10/7/2026, 1:08:42 AM
Qwen3.5-4B ↗	Text generation	BF16	vLLM	512	wbench-http	—	10/7/2026, 2:25:01 AM
Qwen3.5-4B ↗	Text generation	BF16	vLLM	512	Method unknown	—	10/7/2026, 1:09:00 AM
Qwen3.5-9B ↗	Text generation	FP8	openai-compatible	aime2026-test-zero-shot-strict-v1-n30-mt4096-t0	quality-v1	23.33%	10/8/2026, 3:40:59 PM
Qwen3.5-9B ↗	Text generation	FP8	openai-compatible	mmlu-test-zero-shot-strict-v1-n14042-mt256-t0	quality-v1	77.85%	10/8/2026, 3:40:28 PM
Qwen3.5-9B ↗	Text generation	FP8	openai-compatible	gsm8k-test-zero-shot-strict-v1-n1319-mt1024-t0	quality-v1	81.05%	10/8/2026, 3:39:56 PM
Qwen3.5-9B ↗	Text generation	FP8	vllm	2K	vllm-official	—	10/8/2026, 3:39:49 PM
Qwen3.6-27B-AWQ ↗	Text generation	AWQ	vLLM	512	wbench-http	—	10/7/2026, 1:13:02 AM
WeMM-Embedding-2B ↗	Embedding	bf16	vllm	embedding	wbench-http	—	10/8/2026, 10:09:56 AM
whisper-tiny ↗	Speech recognition	fp16	faster-whisper	asr	audio-native-v1	—	10/7/2026, 12:54:21 PM
whisper-tiny ↗	Speech recognition	fp16	faster-whisper	asr	audio-native-v1	—	10/7/2026, 10:25:46 AM
whisper-tiny ↗	Speech recognition	fp16	faster-whisper	asr	audio-native-v1	—	10/7/2026, 10:00:30 AM

Running benchmarks

0Updates automatically

No benchmarks are currently reported as running.

Methodology

Average of error-free runs for the selected recorded configuration. Missing methods and fallback runs are not ranked. Performance index = average / fastest matching hardware × 100. Value index = performance per purchase cost / best comparable value × 100. Quality is separate by dataset and protocol. Prices are dated seller listings, not completed transactions or market averages.

Updated 11/5/2026, 11:06:59 PM

All tasks · Qualified results

embeddinggemma-2

Kokoro-82M

faster-whisper-tiny

whisper-tiny

clef-flash

embeddinggemma-2

Quality by dataset

Original reports

Running benchmarks