← Qwen3 235B A22B Instruct 2507
Qwen3 235B A22B Instruct 2507: 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) | 97.92% | 81.99% | 89.71% | 471ms | 10.7s | 20 tps | 262K | $0.090 / $0.550 | |
Googlegoogle-vertex/us-south1 | Healthy | 100%◆ (best) | 99.86% | 99.61% | 99.82% | 599ms | 3.22s◆ (best) | 46 tps | 262K | $0.220 / $0.880 |
StreamLakestreamlake | Healthy | 100%◆ (best) | 99.29% | 98.63% | 98.81% | 619ms | 16.3s | 57 tps | 128K | $0.210 / $0.840 |
| Healthy | 100%◆ (best) | 97.33% | 90.17% | 93.76% | 605ms | 4.31s | 21 tps | 128K | $0.150 / $0.750 | |
| Healthy | 99.93% | 99.82% | 99.89%◆ (best) | 99.84%◆ (best) | 420ms◆ (best) | 4.51s | 37 tps | 131K | $0.140 / $0.800 | |
| Healthy | 99.81% | 99.46% | 98.59% | 98.55% | 1.26s | 6.11s | 47 tps | 262K | $0.087 / $0.350◆ (best) | |
Alibabaalibaba | Healthy | 99.49% | 98.69% | 97.67% | 99.01% | 545ms | 8.46s | 60 tps◆ (best) | 131K | $0.150 / $0.598 |
| Healthy | 99.26% | 97.35% | 96.65% | 97.79% | 779ms | 44.3s | 39 tps | 131K | $0.090 / $0.580 | |
| Down | 27.27% | 90.03% | 92.33% | 91.48% | 1.33s | 13.8s | 30 tps | 262K | $0.200 / $0.600 | |
| No data | — | — | — | 98.28% | — | — | — | 131K | $0.200 / $0.880 | |
PrimeIntellectprimeintellect | No data | — | — | 95.44% | 95.44% | — | — | — | 262K | $0.250 / $1.00 |
- DeepInfra100%◆ (best)Healthy24h 97.92%
- 7d
- 81.99%
- 30d
- 89.71%
- p50
- 471ms
- p99
- 10.7s
- Throughput
- 20 tps
- Context
- 262K
- $/Mtok
- $0.090 / $0.550
deepinfra/fp8fp8 - Google100%◆ (best)Healthy24h 99.86%
- 7d
- 99.61%
- 30d
- 99.82%
- p50
- 599ms
- p99
- 3.22s◆ (best)
- Throughput
- 46 tps
- Context
- 262K
- $/Mtok
- $0.220 / $0.880
google-vertex/us-south1 - StreamLake100%◆ (best)Healthy24h 99.29%
- 7d
- 98.63%
- 30d
- 98.81%
- p50
- 619ms
- p99
- 16.3s
- Throughput
- 57 tps
- Context
- 128K
- $/Mtok
- $0.210 / $0.840
streamlake - Venice100%◆ (best)Healthy24h 97.33%
- 7d
- 90.17%
- 30d
- 93.76%
- p50
- 605ms
- p99
- 4.31s
- Throughput
- 21 tps
- Context
- 128K
- $/Mtok
- $0.150 / $0.750
venice/fp8fp8 - Parasail99.93%Healthy24h 99.82%
- 7d
- 99.89%◆ (best)
- 30d
- 99.84%◆ (best)
- p50
- 420ms◆ (best)
- p99
- 4.51s
- Throughput
- 37 tps
- Context
- 131K
- $/Mtok
- $0.140 / $0.800
parasail/fp8fp8 - GMICloud99.81%Healthy24h 99.46%
- 7d
- 98.59%
- 30d
- 98.55%
- p50
- 1.26s
- p99
- 6.11s
- Throughput
- 47 tps
- Context
- 262K
- $/Mtok
- $0.087 / $0.350◆ (best)
gmicloud/fp8fp8 - Alibaba99.49%Healthy24h 98.69%
- 7d
- 97.67%
- 30d
- 99.01%
- p50
- 545ms
- p99
- 8.46s
- Throughput
- 60 tps◆ (best)
- Context
- 131K
- $/Mtok
- $0.150 / $0.598
alibaba - Novita99.26%Healthy24h 97.35%
- 7d
- 96.65%
- 30d
- 97.79%
- p50
- 779ms
- p99
- 44.3s
- Throughput
- 39 tps
- Context
- 131K
- $/Mtok
- $0.090 / $0.580
novita/fp8fp8 - Nebius27.27%Down24h 90.03%
- 7d
- 92.33%
- 30d
- 91.48%
- p50
- 1.33s
- p99
- 13.8s
- Throughput
- 30 tps
- Context
- 262K
- $/Mtok
- $0.200 / $0.600
nebius/fp8fp8 - No data24h —
- 7d
- —
- 30d
- 98.28%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 131K
- $/Mtok
- $0.200 / $0.880
atlas-cloud/fp8fp8 - No data24h —
- 7d
- 95.44%
- 30d
- 95.44%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $0.250 / $1.00
primeintellect
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.