Gemma 4 26B A4B : 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Cloudflarecloudflare | Healthy | 100%◆ (best) | 99.86% | 99.81% | 99.55% | 623ms | 12.5s | 49 tps | 256K | $0.100 / $0.300 |
| Healthy | 100%◆ (best) | 99.97% | 99.96% | 99.56% | 350ms | 1.89s | 72 tps | 262K | $0.100 / $0.300 | |
Darkbloomdarkbloom | Healthy | 100%◆ (best) | 99.99% | 99.99%◆ (best) | 99.94%◆ (best) | 1.06s | 5.72s | 44 tps | 131K | $0.042 / $0.220◆ (best) |
| Healthy | 100%◆ (best) | 99.77% | 99.61% | 99.41% | 655ms | 5.12s | 31 tps | 262K | $0.070 / $0.340 | |
Makoramakora | Healthy | 100%◆ (best) | 98.93% | 96.94% | 97.59% | 168ms◆ (best) | 1.27s◆ (best) | 215 tps◆ (best) | 256K | $0.080 / $0.320 |
| Healthy | 100%◆ (best) | 99.81% | 99.79% | 99.49% | 923ms | 3.31s | 43 tps | 262K | $0.130 / $0.400 | |
| Healthy | 100%◆ (best) | 99.07% | 98.89% | 98.86% | 521ms | 3.90s | 94 tps | 262K | $0.130 / $0.400 | |
| Healthy | 100%◆ (best) | 95.67% | 93.59% | 96.43% | 1.27s | 4.54s | 57 tps | 262K | $0.140 / $0.400 | |
| Healthy | 100%◆ (best) | 98.74% | 99.44% | 98.94% | 782ms | 4.57s | 47 tps | 256K | $0.130 / $0.400 | |
| Healthy | 99.97% | 99.93% | 99.92% | 99.31% | 237ms | 1.59s | 81 tps | 262K | $0.076 / $0.255 | |
| Healthy | 99.93% | 99.80% | 99.80% | 99.80% | 551ms | 3.49s | 75 tps | 262K | $0.060 / $0.330 | |
| Healthy | 99.91% | 99.73% | 98.66% | 98.98% | 484ms | 3.56s | 67 tps | 262K | $0.059 / $0.255 | |
Googlegoogle-vertex/global | Healthy | 99.63% | 97.51% | 96.56% | 96.78% | 750ms | 10.2s | 30 tps | 262K | $0.150 / $0.600 |
Rekareka | No data | — | — | 93.26% | 98.59% | — | — | — | 262K | $0.060 / $0.200 |
- Cloudflare100%◆ (best)Healthy24h 99.86%
- 7d
- 99.81%
- 30d
- 99.55%
- p50
- 623ms
- p99
- 12.5s
- Throughput
- 49 tps
- Context
- 256K
- $/Mtok
- $0.100 / $0.300
cloudflare - CoreWeave100%◆ (best)Healthy24h 99.97%
- 7d
- 99.96%
- 30d
- 99.56%
- p50
- 350ms
- p99
- 1.89s
- Throughput
- 72 tps
- Context
- 262K
- $/Mtok
- $0.100 / $0.300
coreweave/bf16bf16 - Darkbloom100%◆ (best)Healthy24h 99.99%
- 7d
- 99.99%◆ (best)
- 30d
- 99.94%◆ (best)
- p50
- 1.06s
- p99
- 5.72s
- Throughput
- 44 tps
- Context
- 131K
- $/Mtok
- $0.042 / $0.220◆ (best)
darkbloom - DeepInfra100%◆ (best)Healthy24h 99.77%
- 7d
- 99.61%
- 30d
- 99.41%
- p50
- 655ms
- p99
- 5.12s
- Throughput
- 31 tps
- Context
- 262K
- $/Mtok
- $0.070 / $0.340
deepinfra/fp8fp8 - Makora100%◆ (best)Healthy24h 98.93%
- 7d
- 96.94%
- 30d
- 97.59%
- p50
- 168ms◆ (best)
- p99
- 1.27s◆ (best)
- Throughput
- 215 tps◆ (best)
- Context
- 256K
- $/Mtok
- $0.080 / $0.320
makora - Novita100%◆ (best)Healthy24h 99.81%
- 7d
- 99.79%
- 30d
- 99.49%
- p50
- 923ms
- p99
- 3.31s
- Throughput
- 43 tps
- Context
- 262K
- $/Mtok
- $0.130 / $0.400
novita/bf16bf16 - Parasail100%◆ (best)Healthy24h 99.07%
- 7d
- 98.89%
- 30d
- 98.86%
- p50
- 521ms
- p99
- 3.90s
- Throughput
- 94 tps
- Context
- 262K
- $/Mtok
- $0.130 / $0.400
parasail/bf16bf16 - SiliconFlow100%◆ (best)Healthy24h 95.67%
- 7d
- 93.59%
- 30d
- 96.43%
- p50
- 1.27s
- p99
- 4.54s
- Throughput
- 57 tps
- Context
- 262K
- $/Mtok
- $0.140 / $0.400
siliconflow/fp8fp8 - Venice100%◆ (best)Healthy24h 98.74%
- 7d
- 99.44%
- 30d
- 98.94%
- p50
- 782ms
- p99
- 4.57s
- Throughput
- 47 tps
- Context
- 256K
- $/Mtok
- $0.130 / $0.400
venice/bf16bf16 - NextBit99.97%Healthy24h 99.93%
- 7d
- 99.92%
- 30d
- 99.31%
- p50
- 237ms
- p99
- 1.59s
- Throughput
- 81 tps
- Context
- 262K
- $/Mtok
- $0.076 / $0.255
nextbit/bf16bf16 - DekaLLM99.93%Healthy24h 99.80%
- 7d
- 99.80%
- 30d
- 99.80%
- p50
- 551ms
- p99
- 3.49s
- Throughput
- 75 tps
- Context
- 262K
- $/Mtok
- $0.060 / $0.330
dekallm/bf16bf16 - Io Net99.91%Healthy24h 99.73%
- 7d
- 98.66%
- 30d
- 98.98%
- p50
- 484ms
- p99
- 3.56s
- Throughput
- 67 tps
- Context
- 262K
- $/Mtok
- $0.059 / $0.255
io-net/bf16bf16 - Google99.63%Healthy24h 97.51%
- 7d
- 96.56%
- 30d
- 96.78%
- p50
- 750ms
- p99
- 10.2s
- Throughput
- 30 tps
- Context
- 262K
- $/Mtok
- $0.150 / $0.600
google-vertex/global - Reka—No data24h —
- 7d
- 93.26%
- 30d
- 98.59%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.060 / $0.200
reka
Measured 4m 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.