Qwen3.5 397B A17B
qwen/qwen3.5-397b-a17b
Compare with Qwen3.5-35B-A3B, Qwen3.5-27B, Qwen3.5-122B-A10B
CompareQwen3.5 397B A17B
| State | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Alibabaalibaba | Healthy | 100% | 98.89% | 97.79% | 97.79% | 990ms | 11.9s | 87 tps | 262K | $0.390 / $2.34 |
| Healthy | 100% | 98.70% | 98.15% | 98.15% | 562ms | 104s | 29 tps | 262K | $0.450 / $3.00 | |
| Healthy | 100% | 97.53% | 97.23% | 97.23% | 638ms | 9.17s | 62 tps | 262K | $0.500 / $3.60 | |
Phalaphala | Healthy | 100% | 95.01% | 96.19% | 96.19% | 2.08s | 35.5s | 50 tps | 262K | $0.550 / $3.50 |
Venicevenice | Healthy | 97.37% | 94.81% | 95.47% | 95.47% | 1.72s | 10.4s | 48 tps | 128K | $0.750 / $4.50 |
StreamLakestreamlake | Healthy | 96.00% | 96.83% | 97.13% | 97.13% | 2.57s | 12.0s | 53 tps | 256K | $0.600 / $3.60 |
| Degraded | 95.74% | 96.71% | 96.03% | 96.03% | 2.29s | 18.0s | 43 tps | 262K | $0.550 / $3.50 | |
DigitalOceandigitalocean | Degraded | — | 97.39% | 94.17% | 94.17% | 1.35s | 13.8s | 6.0 tps | 131K | $0.550 / $3.50 |
| No data | — | 0.00% | 0.00% | 0.00% | — | — | — | 262K | $0.600 / $3.60 | |
Novitanovita | No data | — | 97.16% | 95.59% | 95.59% | 923ms | 7.01s | 8.0 tps | 262K | $0.600 / $3.60 |
- Alibaba100%Healthy24h 98.89%
- 7d
- 97.79%
- 30d
- 97.79%
- p50
- 990ms
- p99
- 11.9s
- Throughput
- 87 tps
- Context
- 262K
- $/Mtok
- $0.390 / $2.34
alibaba - DeepInfra100%Healthy24h 98.70%
- 7d
- 98.15%
- 30d
- 98.15%
- p50
- 562ms
- p99
- 104s
- Throughput
- 29 tps
- Context
- 262K
- $/Mtok
- $0.450 / $3.00
deepinfra/fp8fp8 - Parasail100%Healthy24h 97.53%
- 7d
- 97.23%
- 30d
- 97.23%
- p50
- 638ms
- p99
- 9.17s
- Throughput
- 62 tps
- Context
- 262K
- $/Mtok
- $0.500 / $3.60
parasail/fp8fp8 - Phala100%Healthy24h 95.01%
- 7d
- 96.19%
- 30d
- 96.19%
- p50
- 2.08s
- p99
- 35.5s
- Throughput
- 50 tps
- Context
- 262K
- $/Mtok
- $0.550 / $3.50
phala - Venice97.37%Healthy24h 94.81%
- 7d
- 95.47%
- 30d
- 95.47%
- p50
- 1.72s
- p99
- 10.4s
- Throughput
- 48 tps
- Context
- 128K
- $/Mtok
- $0.750 / $4.50
venice - StreamLake96.00%Healthy24h 96.83%
- 7d
- 97.13%
- 30d
- 97.13%
- p50
- 2.57s
- p99
- 12.0s
- Throughput
- 53 tps
- Context
- 256K
- $/Mtok
- $0.600 / $3.60
streamlake - AtlasCloud95.74%Degraded24h 96.71%
- 7d
- 96.03%
- 30d
- 96.03%
- p50
- 2.29s
- p99
- 18.0s
- Throughput
- 43 tps
- Context
- 262K
- $/Mtok
- $0.550 / $3.50
atlas-cloud/fp8fp8 - Degraded24h 97.39%
- 7d
- 94.17%
- 30d
- 94.17%
- p50
- 1.35s
- p99
- 13.8s
- Throughput
- 6.0 tps
- Context
- 131K
- $/Mtok
- $0.550 / $3.50
digitalocean - No data24h 0.00%
- 7d
- 0.00%
- 30d
- 0.00%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.600 / $3.60
gmicloud/fp8fp8 - No data24h 97.16%
- 7d
- 95.59%
- 30d
- 95.59%
- p50
- 923ms
- p99
- 7.01s
- Throughput
- 8.0 tps
- Context
- 262K
- $/Mtok
- $0.600 / $3.60
novita
Measured 4m ago. Latency and throughput are OpenRouter's rolling 30-minute windows.
Price changes
- Qwen3.5 397B A17BDigitalOceandigitalocean$0.302 / $1.93 → $0.550 / $3.502026-08-21 03:31 UTC
History
2 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.