Qwen3.8 2.4T A95B: 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) | 99.92% | 99.91% | 99.92%◆ (best) | 1.32s | 12.4s | 32 tps | 1.0M | $2.00 / $6.00◆ (best) |
Novitanovita | Healthy | 100%◆ (best) | 99.99% | 99.87% | 99.90% | 4.44s | 42.6s | 33 tps | 1.0M | $2.00 / $6.00◆ (best) |
| Healthy | 99.88% | 99.80% | 99.88% | 99.73% | 4.91s | 32.4s | 30 tps | 1.0M | $2.00 / $6.00◆ (best) | |
| No data | — | 99.99% | 99.75% | 99.48% | 2.44s | 17.0s | 67 tps | 262K | $2.00 / $6.00◆ (best) | |
DigitalOceandigitalocean | No data | — | — | — | — | — | — | — | 262K | $2.00 / $6.00◆ (best) |
Modalmodal | No data | — | 99.21% | 99.20% | 99.74% | 313ms◆ (best) | 2.06s◆ (best) | 173 tps◆ (best) | 1.0M | $2.00 / $6.00◆ (best) |
| No data | — | — | — | — | — | — | — | 1.0M | $2.00 / $6.00◆ (best) | |
Togethertogether | No data | — | 99.95% | 99.99%◆ (best) | 99.90% | 394ms | 3.28s | 156 tps | 1.0M | $2.00 / $6.00◆ (best) |
Venicevenice | No data | — | 99.46% | 98.36% | 99.35% | 1.07s | 76.2s | 69 tps | 262K | $2.00 / $6.00◆ (best) |
- Alibaba100%◆ (best)Healthy24h 99.92%
- 7d
- 99.91%
- 30d
- 99.92%◆ (best)
- p50
- 1.32s
- p99
- 12.4s
- Throughput
- 32 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00◆ (best)
alibaba - Novita100%◆ (best)Healthy24h 99.99%
- 7d
- 99.87%
- 30d
- 99.90%
- p50
- 4.44s
- p99
- 42.6s
- Throughput
- 33 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00◆ (best)
novita - SiliconFlow99.88%Healthy24h 99.80%
- 7d
- 99.88%
- 30d
- 99.73%
- p50
- 4.91s
- p99
- 32.4s
- Throughput
- 30 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00◆ (best)
siliconflow/fp8fp8 - No data24h 99.99%
- 7d
- 99.75%
- 30d
- 99.48%
- p50
- 2.44s
- p99
- 17.0s
- Throughput
- 67 tps
- Context
- 262K
- $/Mtok
- $2.00 / $6.00◆ (best)
deepinfra/fp4fp4 - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $2.00 / $6.00◆ (best)
digitalocean - No data24h 99.21%
- 7d
- 99.20%
- 30d
- 99.74%
- p50
- 313ms◆ (best)
- p99
- 2.06s◆ (best)
- Throughput
- 173 tps◆ (best)
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00◆ (best)
modal - No data24h —
- 7d
- —
- 30d
- —
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00◆ (best)
modal/nvfp4nvfp4 - No data24h 99.95%
- 7d
- 99.99%◆ (best)
- 30d
- 99.90%
- p50
- 394ms
- p99
- 3.28s
- Throughput
- 156 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00◆ (best)
together - No data24h 99.46%
- 7d
- 98.36%
- 30d
- 99.35%
- p50
- 1.07s
- p99
- 76.2s
- Throughput
- 69 tps
- Context
- 262K
- $/Mtok
- $2.00 / $6.00◆ (best)
venice
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.