Nemotron 3.5 Lightning: 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) | 99.95%◆ (best) | 393ms | 672ms◆ (best) | 255 tps◆ (best) | 262K | $0.070 / $0.200 | |
| Healthy | 100%◆ (best) | 99.99% | 99.99% | 99.89% | 275ms◆ (best) | 6.05s | 224 tps | 262K | $0.060 / $0.160 | |
Io Netio-net | Healthy | 100%◆ (best) | 99.68% | 97.88% | 98.49% | 1.03s | 12.0s | 27 tps | 262K | $0.049 / $0.140 |
Phalaphala | Healthy | 100%◆ (best) | 99.81% | 99.51% | 99.17% | 440ms | 1.15s | 160 tps | 262K | $0.070 / $0.200 |
| Degraded | 94.07% | 99.45% | 99.78% | 99.57% | 1.94s | 10.2s | 39 tps | 262K | $0.039 / $0.180◆ (best) | |
| No data | — | — | — | — | — | — | — | 1.0M | $0.100 / $0.250 |
- CoreWeave100%◆ (best)Healthy24h 100%
- 7d
- 100%◆ (best)
- 30d
- 99.95%◆ (best)
- p50
- 393ms
- p99
- 672ms◆ (best)
- Throughput
- 255 tps◆ (best)
- Context
- 262K
- $/Mtok
- $0.070 / $0.200
coreweave/bf16bf16 - DeepInfra100%◆ (best)Healthy24h 99.99%
- 7d
- 99.99%
- 30d
- 99.89%
- p50
- 275ms◆ (best)
- p99
- 6.05s
- Throughput
- 224 tps
- Context
- 262K
- $/Mtok
- $0.060 / $0.160
deepinfra/bf16bf16 - Io Net100%◆ (best)Healthy24h 99.68%
- 7d
- 97.88%
- 30d
- 98.49%
- p50
- 1.03s
- p99
- 12.0s
- Throughput
- 27 tps
- Context
- 262K
- $/Mtok
- $0.049 / $0.140
io-net - Phala100%◆ (best)Healthy24h 99.81%
- 7d
- 99.51%
- 30d
- 99.17%
- p50
- 440ms
- p99
- 1.15s
- Throughput
- 160 tps
- Context
- 262K
- $/Mtok
- $0.070 / $0.200
phala - Darkbloom94.07%Degraded24h 99.45%
- 7d
- 99.78%
- 30d
- 99.57%
- p50
- 1.94s
- p99
- 10.2s
- Throughput
- 39 tps
- Context
- 262K
- $/Mtok
- $0.039 / $0.180◆ (best)
darkbloom/int4int4 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $0.100 / $0.250
venice/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.
Throughput
Median tokens per second.