Artificial Analysis lists Qwen3.5 2B
Artificial Analysis has published a benchmark page for Alibaba's Qwen3.5 2B release, covering two model variants, though the page's text does not include the actual scores.
Independent benchmarking firm Artificial Analysis has added the Qwen3.5 2B release to its model tracker. According to the page, the release covers two models. Each has different intelligence, performance and pricing characteristics.
The page compares the variants on intelligence, cost per Intelligence Index task, time per task, and a set of capability indexes. Those indexes draw on evaluations including AA-Omniscience v1.1, AA-Briefcase v1.1, HLE-Diamond, AutomationBench-AA, GDP.pdf, MLCR-AA, CritPt v2 (Hard Subset), Terminal-Bench 4.0, Terminal-Bench-Science 0.1 and FrontierCode-AA 2.0.
Artificial Analysis says it compares models only within the same class. Non-reasoning models are ranked against other non-reasoning models, and reasoning models are ranked against both kinds. Open-weights models are compared with others of similar size, and anything at 4B parameters or below counts as "Tiny." Proprietary models are grouped by blended price, using a 3:1 input-to-output ratio.
The text of the page does not include any scores, prices or speed figures. It also does not say how the two variants differ, whether the weights are open, or when the models were released. Until those numbers are available, it isn't possible to say where Qwen3.5 2B ranks among other sub-4B models.