Qwen3 VL 30B A3B Instruct
qwen/qwen3-vl-30b-a3b-instruct
Compare with Qwen3 235B A22B Instruct 2507, Qwen3 30B A3B Instruct 2507, Qwen3 Coder 480B A35B
Compare its 4 providers · Incident RSS
Showing only the deepinfra/fp8 provider · show all 4
CompareQwen3 VL 30B A3B Instruct
| State | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Healthy | 99.75% | 99.68% | 99.16% | 99.16% | 436ms | 3.11s | 25 tps | 262K | $0.150 / $0.600 |
- DeepInfra99.75%Healthy24h 99.68%
- 7d
- 99.16%
- 30d
- 99.16%
- p50
- 436ms
- p99
- 3.11s
- Throughput
- 25 tps
- Context
- 262K
- $/Mtok
- $0.150 / $0.600
deepinfra/fp8fp8
Measured 4m ago. Latency and throughput are OpenRouter's rolling 30-minute windows.
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.
Daily availability
Half a year, one cell per UTC day.
Feb
Mar
Apr
May
Jun
Jul
Aug
Mon
Wed
Fri
worst hour ≥ 95%90–95%< 90%no data
Colour by worst endpoint-hour of each UTC day · 3 days with data, 2 with at least one endpoint-hour below 90%.