Step 5 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) | 100% | 100%◆ (best) | 100%◆ (best) | 2.86s | 13.4s | 31 tps | 262K | $1.00 / $2.70◆ (best) | |
| Healthy | 100%◆ (best) | 100% | 100%◆ (best) | 100%◆ (best) | 625ms◆ (best) | 6.67s◆ (best) | 146 tps◆ (best) | 1.0M | $1.00 / $2.70◆ (best) | |
| Healthy | 100%◆ (best) | 98.92% | 99.42% | 99.43% | 5.51s | 82.2s | 34 tps | 1.0M | $1.00 / $2.70◆ (best) | |
| No data | — | 97.97% | 99.75% | 99.75% | 781ms | 17.0s | 123 tps | 1.0M | $1.00 / $2.70◆ (best) | |
Novitanovita | No data | — | — | — | — | — | — | — | 1.0M | $1.00 / $2.70◆ (best) |
- Featherless100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 100%◆ (best)
- p50
- 2.86s
- p99
- 13.4s
- Throughput
- 31 tps
- Context
- 262K
- $/Mtok
- $1.00 / $2.70◆ (best)
featherless/fp4fp4 - InferenceNet100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 100%◆ (best)
- p50
- 625ms◆ (best)
- p99
- 6.67s◆ (best)
- Throughput
- 146 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $1.00 / $2.70◆ (best)
inference-net/fp4fp4 - StepFun100%◆ (best)Healthy24h 98.92%
- 7d
- 99.42%
- 30d
- 99.43%
- p50
- 5.51s
- p99
- 82.2s
- Throughput
- 34 tps
- Context
- 1.0M
- $/Mtok
- $1.00 / $2.70◆ (best)
stepfun/fp8fp8 - No data24h 97.97%
- 7d
- 99.75%
- 30d
- 99.75%
- p50
- 781ms
- p99
- 17.0s
- Throughput
- 123 tps
- Context
- 1.0M
- $/Mtok
- $1.00 / $2.70◆ (best)
deepinfra/fp8fp8 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $1.00 / $2.70◆ (best)
novita
Measured 5m 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.