Qwen3.5-35B-A3B: providers compared
Every endpoint serving this model, side by side. ◆ marks the best value in each column — per model, never fleet-wide.
Current state
| State | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Healthy | 100%◆ (best) | 99.85% | 99.81% | 99.65% | 2.75s | 34.7s | 59 tps | 262K | $0.080 / $0.750◆ (best) | |
| Healthy | 100%◆ (best) | 99.68% | 99.76% | 98.83% | 585ms | 4.58s | 104 tps | 262K | $0.140 / $1.00 | |
| Healthy | 100%◆ (best) | 99.95% | 99.88% | 99.82% | 474ms◆ (best) | 2.95s | 66 tps | 262K | $0.150 / $1.00 | |
Venicevenice | Healthy | 100%◆ (best) | 99.96% | 99.92%◆ (best) | 99.93%◆ (best) | 754ms | 2.43s◆ (best) | 145 tps◆ (best) | 256K | $0.313 / $1.25 |
| No data | — | — | — | — | — | — | — | 262K | $0.140 / $1.00 | |
Alibabaalibaba | No data | — | 91.92% | 95.84% | 96.12% | 488ms | 6.88s | 134 tps | 262K | $0.163 / $1.30 |
| No data | — | 98.46% | 99.12% | 98.53% | 1.47s | 3.91s | 43 tps | 262K | $0.225 / $1.80 | |
| No data | — | — | — | — | — | — | — | 262K | $0.250 / $1.25 | |
| No data | — | 89.85% | 94.91% | 94.56% | 1.51s | 3.50s | 41 tps | 262K | $0.240 / $1.80 |
- Darkbloom100%◆ (best)Healthy24h 99.85%
- 7d
- 99.81%
- 30d
- 99.65%
- p50
- 2.75s
- p99
- 34.7s
- Throughput
- 59 tps
- Context
- 262K
- $/Mtok
- $0.080 / $0.750◆ (best)
darkbloom/fp4fp4 - DeepInfra100%◆ (best)Healthy24h 99.68%
- 7d
- 99.76%
- 30d
- 98.83%
- p50
- 585ms
- p99
- 4.58s
- Throughput
- 104 tps
- Context
- 262K
- $/Mtok
- $0.140 / $1.00
deepinfra/fp8fp8 - Parasail100%◆ (best)Healthy24h 99.95%
- 7d
- 99.88%
- 30d
- 99.82%
- p50
- 474ms◆ (best)
- p99
- 2.95s
- Throughput
- 66 tps
- Context
- 262K
- $/Mtok
- $0.150 / $1.00
parasail/fp8fp8 - Venice100%◆ (best)Healthy24h 99.96%
- 7d
- 99.92%◆ (best)
- 30d
- 99.93%◆ (best)
- p50
- 754ms
- p99
- 2.43s◆ (best)
- Throughput
- 145 tps◆ (best)
- Context
- 256K
- $/Mtok
- $0.313 / $1.25
venice - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.140 / $1.00
akashml/fp8fp8 - No data24h 91.92%
- 7d
- 95.84%
- 30d
- 96.12%
- p50
- 488ms
- p99
- 6.88s
- Throughput
- 134 tps
- Context
- 262K
- $/Mtok
- $0.163 / $1.30
alibaba - No data24h 98.46%
- 7d
- 99.12%
- 30d
- 98.53%
- p50
- 1.47s
- p99
- 3.91s
- Throughput
- 43 tps
- Context
- 262K
- $/Mtok
- $0.225 / $1.80
atlas-cloud/fp8fp8 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.250 / $1.25
coreweave/fp8fp8 - No data24h 89.85%
- 7d
- 94.91%
- 30d
- 94.56%
- p50
- 1.51s
- p99
- 3.50s
- Throughput
- 41 tps
- Context
- 262K
- $/Mtok
- $0.240 / $1.80
siliconflow/fp8fp8
Measured 3m ago. Latency and throughput are OpenRouter's rolling 30-minute windows. ◆ marks the best value in each column among these endpoints.
Price vs latency
One dot per endpoint at the latest round · further left = faster, lower = cheaper · hover a dot for its tag
History
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, averaged over the hour.
Throughput
Median tokens per second.