Unverified50% confidenceBenchmarkExact time
Qwen 3.6 Max在中性模式下端到端准确率仅12.3%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/9/2026
First Seen
Valid until: 10/7/2026
Sources
腾讯DiscoBench：评测LLM搜索代理歧义澄清能力的首个基准
bilibili熊二等兵7/4/2026
Related Claims
UnverifiedQwen3.6 27B版本在Simple QA测试中达到95.7%的准确率76% similarUnverified在终端基准（Terminal Bench）测试中，Qwen 3.6 27B得分约60.7，而Muse Glimmer只有51.7，明显落后76% similarUnverifiedQwen3.5的参数激活率为4.3%，而DeepSeek V3的671B模型激活率为5.5%75% similarUnverifiedQwen3.7 Max的输出质量高度依赖提示词质量，同一模型在不同提示词下的性能差异可高达30%-50%73% similarUnverifiedQwen3.7 Max在通用智能体能力评测中超越了Claude系列模型和前代Qwen3.673% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/387222API
curl https://kongchang.com/api/v1/knowledge/claims/387222MCP
get_claim(id=387222)