Unverified50% confidenceFactExact time
The Jev judge is relatively neutral and also slightly prefers Gemini models, but it cannot give reasons for its scores, which makes it insufficiently stable on these tasks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
UnverifiedThe post's author believes that OpenAI's Luna in its current state is insufficient to compete with Jev.69% similarUnverifiedGemini does not return reasoning traces at the none and low levels, but billing is still calculated based on the tokens actually generated, and hidden thinking tokens are still billed.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011130API
curl https://kongchang.com/api/v1/knowledge/claims/1011130MCP
get_claim(id=1011130)