Qwen3.5-122B-A10B: 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) | 99.64% | 99.22% | 99.59% | 1.58s | 17.3s | 66 tps | 262K | $0.400 / $3.20 | |
| Healthy | 100%◆ (best) | 98.49% | 94.91% | 93.46% | 1.23s | 4.74s | 50 tps | 262K | $0.260 / $2.08◆ (best) | |
Alibabaalibaba | Healthy | 98.81% | 98.91% | 97.55% | 96.16% | 766ms◆ (best) | 2.59s◆ (best) | 77 tps◆ (best) | 262K | $0.260 / $2.08◆ (best) |
| No data | — | 99.00% | 98.20% | 98.01% | 1.35s | 6.43s | 61 tps | 262K | $0.300 / $2.40 | |
| No data | — | — | 99.92%◆ (best) | 99.74%◆ (best) | — | — | — | 262K | $0.290 / $2.40 |
- Novita100%◆ (best)Healthy24h 99.64%
- 7d
- 99.22%
- 30d
- 99.59%
- p50
- 1.58s
- p99
- 17.3s
- Throughput
- 66 tps
- Context
- 262K
- $/Mtok
- $0.400 / $3.20
novita/bf16bf16 - SiliconFlow100%◆ (best)Healthy24h 98.49%
- 7d
- 94.91%
- 30d
- 93.46%
- p50
- 1.23s
- p99
- 4.74s
- Throughput
- 50 tps
- Context
- 262K
- $/Mtok
- $0.260 / $2.08◆ (best)
siliconflow/fp8fp8 - Alibaba98.81%Healthy24h 98.91%
- 7d
- 97.55%
- 30d
- 96.16%
- p50
- 766ms◆ (best)
- p99
- 2.59s◆ (best)
- Throughput
- 77 tps◆ (best)
- Context
- 262K
- $/Mtok
- $0.260 / $2.08◆ (best)
alibaba - No data24h 99.00%
- 7d
- 98.20%
- 30d
- 98.01%
- p50
- 1.35s
- p99
- 6.43s
- Throughput
- 61 tps
- Context
- 262K
- $/Mtok
- $0.300 / $2.40
atlas-cloud/fp8fp8 - No data24h —
- 7d
- 99.92%◆ (best)
- 30d
- 99.74%◆ (best)
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.290 / $2.40
deepinfra/fp4fp4
Measured 2m 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.