Unverified50% confidenceEventExact time
Grok 4.7 was originally scheduled for release last week but was delayed; Musk revealed the reason was that the RL phase may have over-penalized response length, causing the model to give up too early on difficult tasks.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/25/2026
First Seen
Valid until: 12/24/2026
Sources
Related Entities
Related Claims
UnverifiedThe Devin team attributes Grok 4.7's lower aggregate score to the model's tendency to over-scope, proactively introducing additional changes or features beyond what the user requires.70% similarUnverifiedAll three models failed to pass the GRPO stage; while GRPO showed slight improvement in teaching the models arithmetic, general capabilities degraded significantly.63% similarUnverified低精度引入的数值扰动会在RL的多轮迭代中通过策略梯度的乘积形式被放大，可能导致训练崩溃或模型性能退化62% similarUnverifiedPLUG and ILRM were trained only on 16–64 step delays, successfully extrapolating to 4096 steps, achieving 100% accuracy on a one-bit information retention task, while the baseline standard GRU degraded to random-guess level.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/948941API
curl https://kongchang.com/api/v1/knowledge/claims/948941MCP
get_claim(id=948941)