Gemma 3 27B: 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) | 78.31% | 86.79% | 86.79% | 644ms | 2.90s◆ (best) | 20 tps | 110K | $0.100 / $0.300 | |
| Healthy | 99.84% | 99.56% | 98.70% | 98.70% | 1.30s | 6.42s | 15 tps | 131K | $0.080 / $0.160◆ (best) | |
| Healthy | 99.52% | 77.06% | 90.69% | 90.69% | 628ms | 4.76s | 29 tps | 98K | $0.119 / $0.200 | |
| Healthy | 99.23% | 99.38% | 99.26%◆ (best) | 99.26%◆ (best) | 721ms | 5.02s | 12 tps | 131K | $0.080 / $0.450◆ (best) | |
Phalaphala | Down | 33.00% | 93.31% | 94.77% | 94.77% | 567ms◆ (best) | 4.03s | 66 tps◆ (best) | 262K | $0.150 / $0.460 |
- Nebius100%◆ (best)Healthy24h 78.31%
- 7d
- 86.79%
- 30d
- 86.79%
- p50
- 644ms
- p99
- 2.90s◆ (best)
- Throughput
- 20 tps
- Context
- 110K
- $/Mtok
- $0.100 / $0.300
nebius/fp8fp8 - DeepInfra99.84%Healthy24h 99.56%
- 7d
- 98.70%
- 30d
- 98.70%
- p50
- 1.30s
- p99
- 6.42s
- Throughput
- 15 tps
- Context
- 131K
- $/Mtok
- $0.080 / $0.160◆ (best)
deepinfra/fp8fp8 - Novita99.52%Healthy24h 77.06%
- 7d
- 90.69%
- 30d
- 90.69%
- p50
- 628ms
- p99
- 4.76s
- Throughput
- 29 tps
- Context
- 98K
- $/Mtok
- $0.119 / $0.200
novita/bf16bf16 - Parasail99.23%Healthy24h 99.38%
- 7d
- 99.26%◆ (best)
- 30d
- 99.26%◆ (best)
- p50
- 721ms
- p99
- 5.02s
- Throughput
- 12 tps
- Context
- 131K
- $/Mtok
- $0.080 / $0.450◆ (best)
parasail/fp8fp8 - Phala33.00%Down24h 93.31%
- 7d
- 94.77%
- 30d
- 94.77%
- p50
- 567ms◆ (best)
- p99
- 4.03s
- Throughput
- 66 tps◆ (best)
- Context
- 262K
- $/Mtok
- $0.150 / $0.460
phala
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.