‹ PublicAI Index
The LLM benchmark aggregator.
Qwen3.6 27B
Alibaba · 28B · open weights
Strongest in Finance (#51 of 162), weakest in Core abilities (#198 of 199). Above par in 4 of 12 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and GPT-5.5 and ahead of Claude Haiku 4.5 and MiMo V2.5.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Professional51−11.4#86/1941/2
Finance54.8−7.6#51/1621/1
Legal44−21.9#127/1731/1
Agents50.1−16.3#123/2982/6
Knowledge work50.2−20.5#95/2082/2
Coding41.6−24.6#160/1792/6
Agentic coding41.2−25.7#154/1732/5
Reasoning41.9−23.5#163/1951/5
Mathematics37.7−26.1#127/1521/2
Core abilities34.2−31.1#198/1991/4
Language30.5−36.1#59/621/1
Instruction following26−47.8#61/621/1
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Qwen3.6 27B, left for the other.
§ 3 · Sources
Where the numbers come from
4 publications, 16 figures. Every one links to the page it was read from.
GDPval-AA 993
AA-Briefcase 816
LiveBench 1.0 64LiveBench 1.0 · Reasoning 70.3LiveBench 1.0 · Coding 71.8LiveBench 1.0 · Agentic Coding 39.3LiveBench 1.0 · Mathematics 79.9LiveBench 1.0 · Data Analysis 70.4LiveBench 1.0 · Language 63.3LiveBench 1.0 · Instruction Following 53.2
Vals · CaseLaw 53.16%Vals · CorpFin 62.31%Vals · TaxEval 71.26%Vals · MortgageTax 68.28%Vals · SWE-bench Verified 70% (Mini-SWE-agent)Vals · Vibe Code Bench 11.94% (OpenHands)
Badge
[](https://publicai.io/model-index/m/qwen3-6-27b)