Hy3: 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) | 99.81% | 99.83%◆ (best) | 99.57% | 1.34s◆ (best) | 8.92s | 25 tps | 262K | $0.130 / $0.530 | |
| Healthy | 100%◆ (best) | 99.87% | 99.18% | 99.22% | 2.40s | 7.37s | 91 tps | 262K | $0.140 / $0.580 | |
Phalaphala | Healthy | 100%◆ (best) | 99.82% | 99.49% | 99.81% | 2.13s | 20.8s | 93 tps◆ (best) | 262K | $0.150 / $0.640 |
| Healthy | 100%◆ (best) | 99.79% | 99.46% | 99.83%◆ (best) | 3.54s | 15.7s | 62 tps | 262K | $0.132 / $0.528 | |
Novitanovita | Healthy | 98.11% | 99.40% | 98.83% | 99.64% | 4.04s | 23.8s | 61 tps | 262K | $0.140 / $0.580 |
| No data | — | 99.92% | 99.77% | 99.74% | 2.31s | 5.04s◆ (best) | 86 tps | 262K | $0.200 / $0.800 | |
| No data | — | — | — | — | — | — | — | 262K | $0.132 / $0.527 | |
| No data | — | — | — | 96.08% | — | — | — | 262K | $0.105 / $0.435◆ (best) |
- DeepInfra100%◆ (best)Healthy24h 99.81%
- 7d
- 99.83%◆ (best)
- 30d
- 99.57%
- p50
- 1.34s◆ (best)
- p99
- 8.92s
- Throughput
- 25 tps
- Context
- 262K
- $/Mtok
- $0.130 / $0.530
deepinfra/fp4fp4 - GMICloud100%◆ (best)Healthy24h 99.87%
- 7d
- 99.18%
- 30d
- 99.22%
- p50
- 2.40s
- p99
- 7.37s
- Throughput
- 91 tps
- Context
- 262K
- $/Mtok
- $0.140 / $0.580
gmicloud/bf16bf16 - Phala100%◆ (best)Healthy24h 99.82%
- 7d
- 99.49%
- 30d
- 99.81%
- p50
- 2.13s
- p99
- 20.8s
- Throughput
- 93 tps◆ (best)
- Context
- 262K
- $/Mtok
- $0.150 / $0.640
phala - Tencent100%◆ (best)Healthy24h 99.79%
- 7d
- 99.46%
- 30d
- 99.83%◆ (best)
- p50
- 3.54s
- p99
- 15.7s
- Throughput
- 62 tps
- Context
- 262K
- $/Mtok
- $0.132 / $0.528
tencent/fp8fp8 - Novita98.11%Healthy24h 99.40%
- 7d
- 98.83%
- 30d
- 99.64%
- p50
- 4.04s
- p99
- 23.8s
- Throughput
- 61 tps
- Context
- 262K
- $/Mtok
- $0.140 / $0.580
novita - No data24h 99.92%
- 7d
- 99.77%
- 30d
- 99.74%
- p50
- 2.31s
- p99
- 5.04s◆ (best)
- Throughput
- 86 tps
- Context
- 262K
- $/Mtok
- $0.200 / $0.800
atlas-cloud/fp8fp8 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.132 / $0.527
baidu/fp8fp8 - No data24h —
- 7d
- —
- 30d
- 96.08%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.105 / $0.435◆ (best)
deepinfra/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.
Throughput
Median tokens per second.