Hy4 preview: 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) | 95.40% | 99.11% | 98.96% | 1.47s◆ (best) | 2.54s◆ (best) | 66 tps◆ (best) | 1.0M | $0.834 / $2.50 | |
| Healthy | 99.54% | 97.73% | 97.00% | 98.98% | 3.85s | 11.7s | 51 tps | 1.0M | $0.751 / $2.25 | |
| No data | — | 98.17% | 99.75%◆ (best) | 99.75%◆ (best) | 2.24s | 6.16s | 18 tps | 1.0M | $0.834 / $2.50 | |
| No data | — | — | 95.81% | 95.81% | — | — | — | 1.0M | $0.747 / $2.25◆ (best) | |
| Degraded | — | 80.82% | 95.07% | 96.93% | 3.34s | 11.4s | 44 tps | 1.0M | $0.834 / $2.50 |
- SiliconFlow100%◆ (best)Healthy24h 95.40%
- 7d
- 99.11%
- 30d
- 98.96%
- p50
- 1.47s◆ (best)
- p99
- 2.54s◆ (best)
- Throughput
- 66 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $0.834 / $2.50
siliconflow/fp8fp8 - Tencent99.54%Healthy24h 97.73%
- 7d
- 97.00%
- 30d
- 98.98%
- p50
- 3.85s
- p99
- 11.7s
- Throughput
- 51 tps
- Context
- 1.0M
- $/Mtok
- $0.751 / $2.25
tencent/fp8fp8 - No data24h 98.17%
- 7d
- 99.75%◆ (best)
- 30d
- 99.75%◆ (best)
- p50
- 2.24s
- p99
- 6.16s
- Throughput
- 18 tps
- Context
- 1.0M
- $/Mtok
- $0.834 / $2.50
deepinfra/fp8fp8 - No data24h —
- 7d
- 95.81%
- 30d
- 95.81%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.747 / $2.25◆ (best)
inference-net/fp8fp8 - Degraded24h 80.82%
- 7d
- 95.07%
- 30d
- 96.93%
- p50
- 3.34s
- p99
- 11.4s
- Throughput
- 44 tps
- Context
- 1.0M
- $/Mtok
- $0.834 / $2.50
novita/fp8fp8
Measured 1m 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.