Unverified50% confidenceBenchmarkExact time
In testing on the full BEIR SciFact dataset (5,183 documents, 300 queries), Qwen3-Embedding-8B ranked first with 0.7953 nDCG@10, with EmbeddingGemma-300M close behind at 0.7861.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/13/2026
First Seen
Valid until: 1/11/2027
Sources
Related Entities
Related Claims
UnverifiedEmbeddingGemma 2's code retrieval score surpasses that of Qwen3 Embedding at the 0.6B scale, and falls within 2 points of models at the 8B level.70% similarUnverifiedQwen2.5-72B approaches the performance of Llama-3.1-405B across multiple benchmarks.70% similarUnverifiedOn the SciFact dataset (300 queries, 5,183 scientific abstracts, 768 dimensions), the int8+int4 repack model retained 97.9% of the original float32 version's nDCG@10 (0.836 vs 0.854).68% similarUnverifiedQwen3.5-122B-A10B produced 8 refusals in 100 tests, with a KL divergence of 0.0856.67% similarUnverifiedQwen3.8-27B scored 91.1 on OmniDocBench v1.5, the strongest document parsing score among the generative models in this evaluation.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1018769API
curl https://kongchang.com/api/v1/knowledge/claims/1018769MCP
get_claim(id=1018769)