llm.ing

Qwen3.5 35B-A3B

Alibabaopen weightsOpenRouter ↗Hugging Face ↗

Qwen vision-language model for visual reasoning, documents, and agent tasks

Benchmarks

BenchmarkCategoryScoreVariantProvenanceSourceObserved
AA Agentic Indexagentic11.8—mirroropenrouterJul 23, 2026
AA Coding Indexcoding37.0—mirroropenrouterJul 23, 2026
AA Intelligence Indexintelligence19.3—independentaaSep 9, 2026
19.3—mirroropenrouterOct 5, 2026
LMArena Textpreference1395—crowdlmarenaNov 9, 2026
LMArena WebDevpreference1253—crowdlmarenaNov 11, 2026
Median output speedspeed145 tok/s—independentaaSep 28, 2026
Median time to first tokenspeed2.15s—independentaaSep 28, 2026

Includes index data from Artificial Analysis.

Providers

Live operational stats per hosting endpoint, via OpenRouter.

ProviderQuantContext$/1M in · outUptime 24hLatencyThroughput
Alibaba—262K$0.163 · $1.3100.00%——
AtlasCloudfp8262K$0.225 · $1.893.10%——
Darkbloomfp4262K$0.065 · $0.7599.75%——
DeepInfrafp8262K$0.14 · $1100.00%——
DeepInfrafp8262K$0.112 · $0.80.00%——
Parasailfp8262K$0.15 · $199.90%——
SiliconFlowfp8262K$0.24 · $1.897.30%——
Venice—256K$0.15 · $199.91%——