Unverified50% confidenceBenchmarkExact time
The research team conducted zero-shot evaluations of off-the-shelf models on 15 long-context tasks, with test subjects including Gemma-4-31B and Qwen-3.6-27B.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/6/2026
First Seen
Valid until: 12/5/2026
Sources
Related Entities
Related Claims
UnverifiedEarly testing on the Opus model showed an exact string recognition rate of 0 out of 15, whereas after the release of Fable 5 the same test yielded 13 out of 15.65% similarUnverifiedThis test generated a total of 147 model calls.65% similarUnverifiednomic-embed-text嵌入模型有137M参数，all-MiniLM-L6-v2有22M参数，BGE-small为24M-109M参数64% similarUnverifiedReddit社区的大规模基准测试基于BeeLlama.cpp v0.4.0，在Qwen 3.6 27B和Gemma 4 31B两款模型上累计跑了413组配置对比，其中Qwen模型238组、Gemma模型175组63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/865959API
curl https://kongchang.com/api/v1/knowledge/claims/865959MCP
get_claim(id=865959)