Qwen3 VL 30B A3B Instruct: 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) | 99.66% | 99.75%◆ (best) | 91.35% | 1.57s | 6.09s | 19 tps | 262K | $0.150 / $0.600 | |
Alibabaalibaba | Healthy | 99.78% | 99.37% | 99.39% | 99.75%◆ (best) | 575ms◆ (best) | 2.80s◆ (best) | 64 tps◆ (best) | 131K | $0.130 / $0.520 |
| Healthy | 95.96% | 97.64% | 98.68% | 97.74% | 1.27s | 6.22s | 52 tps | 131K | $0.200 / $0.700 | |
| Degraded | 93.97% | 87.14% | 69.50% | 83.27% | 1.92s | 6.49s | 19 tps | 262K | $0.290 / $1.00 | |
| No data | — | — | — | — | — | — | — | 131K | $0.090 / $0.400◆ (best) | |
PrimeIntellectprimeintellect | No data | — | — | 99.09% | 99.09% | — | — | — | 262K | $0.200 / $0.700 |
- DeepInfra100%◆ (best)Healthy24h 99.66%
- 7d
- 99.75%◆ (best)
- 30d
- 91.35%
- p50
- 1.57s
- p99
- 6.09s
- Throughput
- 19 tps
- Context
- 262K
- $/Mtok
- $0.150 / $0.600
deepinfra/fp8fp8 - Alibaba99.78%Healthy24h 99.37%
- 7d
- 99.39%
- 30d
- 99.75%◆ (best)
- p50
- 575ms◆ (best)
- p99
- 2.80s◆ (best)
- Throughput
- 64 tps◆ (best)
- Context
- 131K
- $/Mtok
- $0.130 / $0.520
alibaba - Novita95.96%Healthy24h 97.64%
- 7d
- 98.68%
- 30d
- 97.74%
- p50
- 1.27s
- p99
- 6.22s
- Throughput
- 52 tps
- Context
- 131K
- $/Mtok
- $0.200 / $0.700
novita/bf16bf16 - SiliconFlow93.97%Degraded24h 87.14%
- 7d
- 69.50%
- 30d
- 83.27%
- p50
- 1.92s
- p99
- 6.49s
- Throughput
- 19 tps
- Context
- 262K
- $/Mtok
- $0.290 / $1.00
siliconflow/fp8fp8 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 131K
- $/Mtok
- $0.090 / $0.400◆ (best)
darkbloom/fp4fp4 - No data24h —
- 7d
- 99.09%
- 30d
- 99.09%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.200 / $0.700
primeintellect
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, averaged over the hour.
Throughput
Median tokens per second.