NEWMEN
← back to all models

Qwen: Qwen3 VL 235B A22B Instruct

qwen/qwen3-vl-235b-a22b-instruct

Alibaba (Qwen)open-weight262K context6 providersIntelligence 71.0· est.

Cheapest provider

$0.20 / 1M

DeepInfra

Fastest provider (p95)

59 tok/s

Venice

Intelligence (estimated)

71.0

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.

ProviderQuantInput $/1MOutput $/1MLatency p50 / p95Throughput p50 / p95Uptime 30mSuccess
DeepInfraq8· fp8$0.2000$0.88001274ms / 5028ms26 / 37 tok/s96.71%86.4%
Parasailq8· fp8$0.2100$1.9000844ms / 3537ms20 / 52 tok/s99.54%99.5%
Veniceq8· fp8$0.2500$1.50002166ms / 4950ms28 / 59 tok/s96.81%77.0%
Alibabaundisclosed$0.2600$1.0400835ms / 2199ms39 / 53 tok/s100.00%100.0%
AtlasCloudq8· fp8$0.3000$1.50002191ms / 4703ms6 / 19 tok/s81.64%81.1%
Novitafull· bf16$0.3000$1.50001496ms / 6019ms19 / 46 tok/s80.86%79.6%

“—” 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.

Want a curated score? File a benchmark report and we’ll add it to the next sync.

How Atlas mode sources Qwen: Qwen3 VL 235B A22B Instruct

  • Strict mode — pin Qwen: Qwen3 VL 235B A22B Instruct exactly and we pass it straight through, sourced from the cheapest provider above. The same model, no substitutions — currently DeepInfra at $0.20/1M.
  • 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.