Unverified50% confidenceFactExact time
The performance metrics claimed by Nari come primarily from the project team's own statements, and independent benchmark data and side-by-side comparisons with mainstream solutions such as Whisper are still lacking
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/14/2026
First Seen
Valid until: 12/13/2026
Sources
Related Entities
Related Claims
UnverifiedThe user pointed out that these models achieve high scores on benchmarks yet frequently fail at the simplest tasks.69% similarUnverifiedbenchmark分数与实际开发体验之间存在显著落差，模型可能在标准测试中领先却在真实项目中表现平平68% similarPartially VerifiedBenchmark contamination refers to a model having already seen test questions during training, resulting in inflated scores that fail to reflect true generalization ability.65% similarUnverifiedNsight Systems通过低开销的系统级追踪技术采集数据，数据采集对被测程序的性能影响通常低于5%64% similarUnverifiedFor small models with limited parameter counts, data quality can have an even greater impact on final performance than data quantity.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/921025API
curl https://kongchang.com/api/v1/knowledge/claims/921025MCP
get_claim(id=921025)