Unverified50% confidenceFactExact time
Qwen3 27B uses linear attention in 48 of its 64 layers, with the remaining 16 layers paired with standard gated attention.
1
Sources
50%
Confidence
Long-term
Relevance
9/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedQwen3 27B scored 76 points on the visual reasoning leaderboard.68% similarUnverifiedQwen3 27B is in the same tier as the DeepSeek V-Flash series on the overall intelligence leaderboard, with scores in the 52–53 range.65% similarUnverifiedQwen-2 introduced grouped query attention (GQA), a longer context window, and an improved tokenizer, with parameter scales ranging from 0.5B to 72B.64% similarUnverifiedQwen3.8-27B enables Thinking mode by default, with reasoning depth available in three levels: Low, Medium, and X-High.63% similarUnverifiedGlimmer's intelligence index score of 35 surpasses Gemma 2 27B (30) and nearly matches Qwen 2.5 (36), with Qwen 2.5 72B topping the leaderboard at 38.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/852383API
curl https://kongchang.com/api/v1/knowledge/claims/852383MCP
get_claim(id=852383)