Unverified50% confidenceBenchmarkExact time
The random-position null baseline for Qwen3-4B was only 11.3%, and the upper bound for directly patching lower-level prompt facts was 83%.
1
Sources
50%
Confidence
Long-term
Relevance
9/24/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedQwen3-1.7B's overall causal faithfulness is only 54.8%, far below that of Qwen3-4B.78% similarUnverifiedQwen 3.6 Max has an end-to-end accuracy of only 12.3% in neutral mode73% similarUnverifiedQwen3 27B在OpenAI HumanEval 164道Python题上取得85%通过率和99%应答率70% similarUnverifiedOn the blogger's overall leaderboard, Qwen3 Max 0902 ranks just below Fable 5.1 (92.5%) and GLM 5.3 (91.25%).68% similarUnverifiedThe UP host stated that the Qwen3 coding model scored approximately 61.7 on SWE-Bench Pro, while the top-tier closed-source model used as a reference scored 53.4 (this data contains inconsistencies and is for reference only).67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/946172API
curl https://kongchang.com/api/v1/knowledge/claims/946172MCP
get_claim(id=946172)