Qwen: Qwen3 VL 8B Instruct
qwen/qwen3-vl-8b-instruct
Cheapest provider
$0.08 / 1M
AtlasCloud
Fastest provider (p95)
120 tok/s
Parasail
Intelligence (estimated)
58.0
Family + generation + popularity + price tier
Per-provider performance
Latency / throughput / uptime / price measured across providers over the last 30 minutes of live traffic. This is what proves “sourced cheapest” — Atlas mode draws on these per call to serve the cheapest path that holds quality.
| Provider | Quant | Input $/1M | Output $/1M | Latency p50 / p95 | Throughput p50 / p95 | Uptime 30m | Success |
|---|---|---|---|---|---|---|---|
| AtlasCloud | q8· fp8 | $0.0800 | $0.5000 | 647ms / 1268ms | 1 / 8 tok/s | 99.29% | 99.4% |
| Novita | q8· fp8 | $0.0800 | $0.5000 | 788ms / 1629ms | 30 / 49 tok/s | 99.25% | 99.3% |
| Alibaba | undisclosed | $0.1170 | $0.4550 | 436ms / 1120ms | 15 / 111.1 tok/s | 99.95% | 99.3% |
| Parasail | full· bf16 | $0.2500 | $0.7500 | 481ms / 2181ms | 18 / 120 tok/s | 99.92% | 94.7% |
“—” means live telemetry hasn’t accumulated enough recent traffic for that endpoint. “undisclosed” means the provider serves the model but doesn’t expose the quantization label (typically running fp8 / int8 internally).
Intelligence estimate
No public benchmark numbers indexed for this model yet, so the leaderboard score is derived from the catalogue data we sync: model family standing, generation, and live usage rank — applied identically to open- and closed-weight models. The heuristic ceiling sits below confirmed-benchmark frontier models so curated rankings stay clearly on top.
- ·qwen sm family base
Want a curated score? File a benchmark report and we’ll add it to the next sync.
How Atlas mode sources Qwen: Qwen3 VL 8B Instruct
- Strict mode — pin Qwen: Qwen3 VL 8B Instruct exactly and we pass it straight through, sourced from the cheapest provider above. The same model, no substitutions — currently AtlasCloud at $0.08/1M.
- Atlas mode — the default. Each call is auto-optimized for the cheapest path that holds quality, at least 5% off going direct from call one and climbing as it ramps. You always see which model served the call and exactly what you saved — thumbs-down anything you don’t like for a full refund.