The consecutive failures of Opus 4.7 and 4.8 are puzzling
TL;DR · AI Summary
The updates to Opus 4.7 and 4.8 did not meet expectations, with higher prices but no improvement in performance, or even a decline. The author points out that the Claude model has been stagnant for four months and questions why progress is still slow despite having mastered training methods and possessing advanced tools like Mythos.
Key Takeaways
- Upgrades to Opus 4.7 and 4.8 were disappointing, with increased prices but no si
- The Claude model has been without updates for over four months
- Despite having the capability to train models and using advanced tools like Myth
Outline
Jump quickly between sections.
Recent versions of Opus have underperformed expectations
Higher prices did not lead to better performance, with some regressions observed
Noticed that the Claude model has not been updated for four months
Even with mastery of training methods and use of advanced tools, progress remains sluggish
Current iteration cycles are too long to meet demands
Mindmap
See how the topics connect at a glance.
查看大纲文本（无障碍 / 无 JS 友好）
- 模型更新困境
- Opus版本问题
- 性能不升反降
- Claude停更现象
- 持续四个月无新版本
- 技术瓶颈
- 训练方法与工具不足
Highlights
Key sentences worth saving and sharing.
Higher prices for Opus 4.7 and 4.8 did not result in improved effectiveness
Claude has been stagnant for over four months
Despite advancements like Mythos, model progress remains slow
Orange AI on X: "Opus 4.7, 4.8接连失败令人费解 价格更贵，效果无提升，甚至负提升 看看日历，突然意识到 Claude已经停滞了4个月 即便是掌握了模型训练的方法，即便内部已经有了Mythos这样的开发利器 模型的进步还是没有太多加速，依然半年一次大更新？ https://t.co/eDTgiMmogg" / X
Show translation
Opus 4.7、4.8的接连失败令人费解 价格更贵，效果无提升，甚至负提升 看看日历，突然意识到 Claude已经停滞了4个月 即便是掌握了模型训练的方法，即便内部已经有了Mythos这样的开发利器 模型的进步还是没有太多加速，依然半年一次大更新？
2