Unverified50% confidenceBenchmarkExact time
Both Jev and Drex support up to 255 options; Jev reaches 79.2% on the full Banking77 test (77 labels); Laya's performance degrades beyond about 20 options under default settings, reaching only 0.425 on Banking77.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/19/2026
First Seen
Valid until: 1/17/2027
Sources
Related Entities
Related Claims
UnverifiedAfter fine-tuning, Laya reaches 0.766 on the same benchmark, versus the third-party-published Jev figure of 0.727; Drex offers fine-tuning as a vendor service.70% similarUnverifiedOn six public text decision benchmarks, the six-group mean of the JEV-27B / JEV-27B-VL rows is 84.07%, higher than the 83.85% of the TypeSafe Jev 1.13 hosted API and the 77.70% of NeoHorse-Jev-4B.68% similarUnverifiedJev can only handle 255 candidate options at a time.61% similarUnverifiedLaya has only 421 million parameters and claims to surpass the original architecture on all Jev benchmarks.60% similarUnverifiedpplx-decider-v1-27b achieved an average score of 85.7% across 11 benchmarks, outperforming the comparison target Jev (official data).59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1065252API
curl https://kongchang.com/api/v1/knowledge/claims/1065252MCP
get_claim(id=1065252)