Qwen3.8 2.4T A95B
qwen/qwen3.8-2.4t-a95b
Compare with Qwen3 235B A22B Instruct 2507, Qwen3.5 397B A17B, Qwen3.6 35B A3B
| Provider | State | Uptime 5m | 24h | p50 latency | p99 | Throughput | Context | $/Mtok |
|---|---|---|---|---|---|---|---|---|
DeepInfradeepinfra/fp4fp4 | Degraded | 100% | 95.55% | 1.78s | 11.6s | 55 tps | 262K | $2.00 / $6.00 |
Modalmodal/nvfp4nvfp4 | Healthy | 100% | 99.51% | 1.19s | 9.52s | 76 tps | 1.0M | $2.00 / $6.00 |
Alibabaalibaba | No data | — | 99.94% | 1.49s | 2.37s | 32 tps | 1.0M | $2.00 / $6.00 |
DigitalOceandigitalocean | No data | — | 98.28% | — | — | — | 262K | $2.00 / $6.00 |
SiliconFlowsiliconflow/fp8fp8 | No data | — | 99.48% | 2.97s | 11.9s | 33 tps | 1.0M | $2.00 / $6.00 |
Togethertogether | No data | — | 99.86% | 4.47s | 13.3s | 76 tps | 1.0M | $2.50 / $6.25 |
Venicevenice | No data | — | 98.37% | 1.35s | 38.9s | 70 tps | 262K | $2.50 / $7.50 |
- DeepInfra100%Degraded24h 95.55%
- p50
- 1.78s
- p99
- 11.6s
- Throughput
- 55 tps
- Context
- 262K
- $/Mtok
- $2.00 / $6.00
deepinfra/fp4fp4 - Modal100%Healthy24h 99.51%
- p50
- 1.19s
- p99
- 9.52s
- Throughput
- 76 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00
modal/nvfp4nvfp4 - Alibaba—No data24h 99.94%
- p50
- 1.49s
- p99
- 2.37s
- Throughput
- 32 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00
alibaba - DigitalOcean—No data24h 98.28%
- p50
- —
- p99
- —
- Throughput
- —
- Context
- 262K
- $/Mtok
- $2.00 / $6.00
digitalocean - SiliconFlow—No data24h 99.48%
- p50
- 2.97s
- p99
- 11.9s
- Throughput
- 33 tps
- Context
- 1.0M
- $/Mtok
- $2.00 / $6.00
siliconflow/fp8fp8 - Together—No data24h 99.86%
- p50
- 4.47s
- p99
- 13.3s
- Throughput
- 76 tps
- Context
- 1.0M
- $/Mtok
- $2.50 / $6.25
together - Venice—No data24h 98.37%
- p50
- 1.35s
- p99
- 38.9s
- Throughput
- 70 tps
- Context
- 262K
- $/Mtok
- $2.50 / $7.50
venice
Measured 3m ago. Latency and throughput are OpenRouter's rolling 30-minute windows.
History
5 endpoints reporting no data are omitted from the charts.
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 and p99 are OpenRouter's rolling 30-minute windows.
Throughput
Median tokens per second.