Qwen3 Next 80B 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 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Alibabaalibaba | Healthy | 100%◆ (best) | 99.86% | 99.78% | 99.92% | 403ms◆ (best) | 911ms◆ (best) | 95 tps | 131K | $0.098 / $0.780 |
| Healthy | 100%◆ (best) | 99.79% | 99.55% | 99.37% | 529ms | 6.02s | 137 tps | 262K | $0.090 / $1.10◆ (best) | |
Googlegoogle-vertex/global | Healthy | 100%◆ (best) | 99.91% | 99.81% | 99.88% | 603ms | 2.35s | 155 tps◆ (best) | 262K | $0.150 / $1.20 |
| Down | 100%◆ (best) | 84.97% | 88.63% | 93.59% | 1.39s | 74.3s | 76 tps | 131K | $0.150 / $1.50 | |
| Healthy | 100%◆ (best) | 99.96% | 99.96%◆ (best) | 99.95%◆ (best) | 1.15s | 4.83s | 98 tps | 262K | $0.100 / $1.10 |
- Alibaba100%◆ (best)Healthy24h 99.86%
- 7d
- 99.78%
- 30d
- 99.92%
- p50
- 403ms◆ (best)
- p99
- 911ms◆ (best)
- Throughput
- 95 tps
- Context
- 131K
- $/Mtok
- $0.098 / $0.780
alibaba - DeepInfra100%◆ (best)Healthy24h 99.79%
- 7d
- 99.55%
- 30d
- 99.37%
- p50
- 529ms
- p99
- 6.02s
- Throughput
- 137 tps
- Context
- 262K
- $/Mtok
- $0.090 / $1.10◆ (best)
deepinfra/fp8fp8 - Google100%◆ (best)Healthy24h 99.91%
- 7d
- 99.81%
- 30d
- 99.88%
- p50
- 603ms
- p99
- 2.35s
- Throughput
- 155 tps◆ (best)
- Context
- 262K
- $/Mtok
- $0.150 / $1.20
google-vertex/global - Novita100%◆ (best)Down24h 84.97%
- 7d
- 88.63%
- 30d
- 93.59%
- p50
- 1.39s
- p99
- 74.3s
- Throughput
- 76 tps
- Context
- 131K
- $/Mtok
- $0.150 / $1.50
novita/bf16bf16 - Parasail100%◆ (best)Healthy24h 99.96%
- 7d
- 99.96%◆ (best)
- 30d
- 99.95%◆ (best)
- p50
- 1.15s
- p99
- 4.83s
- Throughput
- 98 tps
- Context
- 262K
- $/Mtok
- $0.100 / $1.10
parasail/fp8fp8
Measured 1m 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.