Unverified50% confidenceReleaseExact time
The model trained with the PPRS method is Tev-27B.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/14/2026
First Seen
Valid until: 1/12/2027
Sources
Related Entities
Related Claims
UnverifiedTev-27B was trained entirely with PPRS on the computational and mathematical sciences domains of ALE (Agents' Last Exam), matching GPT-6 Astra and Claude Opus 5 on the Pass@5 metric of ALE-CM.70% similarUnverifiedTev-27B achieves Claude Opus 4.8-level performance on the medium-horizon task TB 2.1.61% similarUnverifiedIn tests on 2,456 held-out prompts, increasing the sampling steps from 20 to 50 improved FID only from 27.0 to 26.6 (a gain of 0.4), while removing the shift caused FID to rise from 27.0 to 27.3 and FD-DINOv2 to rise from 218 to 228.59% similarUnverifiedThe user ultimately settled on a 27B uncensored model, having previously iterated through 27B and 35B MoE models in the Qwen series.59% similarUnverifiedQwen 2.5 Coder, with only 32B parameters, matched the larger general-purpose model GPT-4o, indicating that code-specialized training has a punch-above-its-weight effect on complex array shape and broadcasting errors.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1028129API
curl https://kongchang.com/api/v1/knowledge/claims/1028129MCP
get_claim(id=1028129)