Unverified50% confidencePredictionExact time
Rumor has it that V4.1 will be built on a brand-new pre-training foundation.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedLoRA fine-tuning based on pre-trained models is the most realistic technical path for individual developers.70% similarUnverifiedv6 was trained from scratch using an entirely new dataset that does not contain the same data used by previous models.68% similarVerifiedThe GLM series originally adopted an autoregressive blank-filling pre-training framework that accommodates both understanding and generation capabilities.67% similarPartially VerifiedA pre-trained model refers to a model that has been trained in advance on large-scale datasets, which users can directly fine-tune or call to complete specific tasks.67% similarUnverifiedThe system supports model management functionality, allowing the newly trained V2.0 (10-epoch training version) model to be uploaded and set as the default66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/904548API
curl https://kongchang.com/api/v1/knowledge/claims/904548MCP
get_claim(id=904548)