Qwen3.6 27B
qwen/qwen3.6-27b
Compare with Qwen3.6 35B A3B, Qwen-Plus, Qwen3.6 Flash
CompareQwen3.6 27B
| State | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Healthy | 100% | 99.93% | 99.91% | 99.91% | 436ms | 5.52s | 52 tps | 262K | $0.600 / $3.60 | |
| Healthy | 100% | 99.40% | 97.33% | 97.33% | 1.45s | 15.5s | 26 tps | 262K | $0.320 / $3.20 | |
| Healthy | 100% | 96.35% | 95.45% | 95.45% | 970ms | 14.4s | 23 tps | 256K | $0.325 / $3.25 | |
| Healthy | 98.96% | 93.63% | 70.73% | 70.73% | 2.38s | 22.0s | 17 tps | 262K | $0.300 / $2.00 | |
Alibabaalibaba | No data | — | 99.75% | 99.78% | 99.78% | 2.48s | 38.7s | 39 tps | 262K | $0.450 / $2.70 |
Phalaphala | No data | — | 95.21% | 97.12% | 97.12% | 1.38s | 14.3s | 38 tps | 262K | $0.320 / $2.70 |
| No data | — | 93.36% | 92.59% | 92.59% | 1.72s | 15.8s | 32 tps | 262K | $0.300 / $3.20 |
- CoreWeave100%Healthy24h 99.93%
- 7d
- 99.91%
- 30d
- 99.91%
- p50
- 436ms
- p99
- 5.52s
- Throughput
- 52 tps
- Context
- 262K
- $/Mtok
- $0.600 / $3.60
coreweave/fp8fp8 - DeepInfra100%Healthy24h 99.40%
- 7d
- 97.33%
- 30d
- 97.33%
- p50
- 1.45s
- p99
- 15.5s
- Throughput
- 26 tps
- Context
- 262K
- $/Mtok
- $0.320 / $3.20
deepinfra/fp8fp8 - Venice100%Healthy24h 96.35%
- 7d
- 95.45%
- 30d
- 95.45%
- p50
- 970ms
- p99
- 14.4s
- Throughput
- 23 tps
- Context
- 256K
- $/Mtok
- $0.325 / $3.25
venice/fp8fp8 - Chutes98.96%Healthy24h 93.63%
- 7d
- 70.73%
- 30d
- 70.73%
- p50
- 2.38s
- p99
- 22.0s
- Throughput
- 17 tps
- Context
- 262K
- $/Mtok
- $0.300 / $2.00
chutes/fp8fp8 - No data24h 99.75%
- 7d
- 99.78%
- 30d
- 99.78%
- p50
- 2.48s
- p99
- 38.7s
- Throughput
- 39 tps
- Context
- 262K
- $/Mtok
- $0.450 / $2.70
alibaba - No data24h 95.21%
- 7d
- 97.12%
- 30d
- 97.12%
- p50
- 1.38s
- p99
- 14.3s
- Throughput
- 38 tps
- Context
- 262K
- $/Mtok
- $0.320 / $2.70
phala - No data24h 93.36%
- 7d
- 92.59%
- 30d
- 92.59%
- p50
- 1.72s
- p99
- 15.8s
- Throughput
- 32 tps
- Context
- 262K
- $/Mtok
- $0.300 / $3.20
siliconflow/fp8fp8
Measured 2m ago. Latency and throughput are OpenRouter's rolling 30-minute windows.
History
3 endpoints reporting no data are omitted from the charts.
Availability
Axis starts at 90% so small differences stay visible; it expands to 0 when an endpoint falls below.
Time to first token
Logarithmic axis — the fleet spans two orders of magnitude. P50 is OpenRouter's rolling 30-minute window.
Throughput
Median tokens per second.