Unverified50% confidenceBenchmarkExact time
Qwen3 27B's score on the long-horizon autonomous coding metric DeepSWE jumped from 13.3 in the previous generation to 42.2.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/6/2026
First Seen
Valid until: 12/5/2026
Sources
Related Entities
Related Claims
UnverifiedQwen3 27B features 64 layers, a hidden dimension of 5120, and a native context length of 262,144 tokens.70% similarUnverifiedQwen3 27B scored 90.3 on LiveCodeBench V6, slightly surpassing Opus 4.6 Max's score of 88.8.69% similarUnverifiedQwen3.6 27B supports a context length of 262K, and the 34B 80-layer version maintains the same context length unchanged68% similarVerifiedQwen3-27B natively supports a long context of approximately 260,000 tokens, which can be extended to the 1,000,000-token scale through technologies such as YaRN.66% similarUnverifiedThe Qwen3.6 34B 80-layer model was created by community developers who expanded the 64-layer Transformer network of Qwen3.6 27B to 80 layers, growing the parameter count from 27B to 34B, an increase of approximately 26%66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/863357API
curl https://kongchang.com/api/v1/knowledge/claims/863357MCP
get_claim(id=863357)