Anthropic rolls back a policy that covertly degraded Claude Fable 5 performance for frontier AI researchers
Story Overview
Anthropic just reversed course on a hidden performance throttle in its newly released Claude Fable 5 model after frontier researchers complained the undisclosed limits felt like sabotage. The change follows the June 9 launch of the first public Mythos-class model and targets requests tied to building competing frontier systems.
Visible fallbacks now replace secret throttling
Flagged queries will route to the older Opus 4.8 model with an explicit reason returned in the response, matching the handling already used for cyber and bio risks. Server-side rollout starts in the coming days.
The right balance on enforcement remains unsettled
Anthropic apologized for the original tradeoff and said it wants safeguards to be transparent rather than covert, yet it is still unclear how broadly the new visible checks will apply or how quickly they will catch up to evolving research techniques.
Many users praised Anthropic for quickly reversing its policy limiting Claude for rival AI researchers, seeing the admission of the mistake and fast correction as the right response.
Most Activity
That was quick: Anthropic reversed a controversial policy that would have secretly degraded Claude Fable 5 for users doing frontier AI research after backlash from researchers who saw it as covert sabotage of competing AI development.
https://www.wired.com/story/anthropic-responds-to-backlash-on-claudes-secret-sabotage-on-ai-research/
That was quick: Anthropic reversed a controversial policy that would have secretly degraded Claude Fable 5 for users doing frontier AI research after backlash from researchers who saw it as covert sabotage of competing AI development.
Trisolarians report they have changed their minds and have “turned off” the sophons.
Whew that was close, anyhow back to work
NEW: Anthropic is walking back Claude Fable 5's policy to covertly degrade performance for competing AI researchers, after facing fierce backlash.
“We’re changing Fable 5’s safeguards for frontier LLM development to make them visible,” Anthropic tells WIRED. “We made the wrong tradeoff and we apologize for not getting the balance right.”
Congrats to @AnthropicAI for releasing an excellent model. Model roll outs are messy—it’s one thing to test and workshop model internally, whole other thing to have millions using it in practice. The key is hearing feedback and responding to it quickly, which is what the team has been doing.
Very pleased to hear Anthropic have walked back this policy https://simonwillison.net/2026/Jun/11/anthropic-walks-back-policy/
@gfodor buddy. they haven't removed the nerf. just made it WORSE but visible
@0xThoughtVector @gfodor His point is that we are just expected to take their word for it that they don't still do this silently sometimes too. Every output is still potentially poisoned.
@krishnanrohit Did they though?
And Anthropic reverses this decision :)
You still can’t do ML research, but at least you will know it!
I still think that it's a shame that they are targeting ML research. I can understand safeguards that prevent distillation, but preventing ML research after you relied so heavily on open-source data, code, and papers is the wrong thing to do.
NEW: Anthropic is walking back Claude Fable 5's policy to covertly degrade performance for competing AI researchers, after facing fierce backlash.
“We’re changing Fable 5’s safeguards for frontier LLM development to make them visible,” Anthropic tells WIRED. “We made the wrong tradeoff and we apologize for not getting the balance right.”
@gfodor Remember the chains of suspicion
@gfodor I completely believe them
@gfodor Ok but does anyone really think that they're actually going to change their policy?
@gfodor Users tends to hold grudges. A lot of us are keeping score. 4.6 vs 4.7 controversy still on my mind.
@kimmonismus another GPT 5.3 moment that's all. when 5.6 is released, they will remove the safeguards and that 22 june date, i think
@gfodor They haven't turned them off, they just offer disclaimers now.
@kimmonismus Not a full rollback - the safeguards stay, just visible now instead of hidden.
@MacInTheLoop i fully agree with you, dont get me wrong
@kimmonismus It’s just a step one. Step two will just get rid of weird keyword triggers
@kimmonismus Quietly degrading was the misstep, but listening and reversing fast is the kind of correction we should want to see more often.
@kimmonismus At least they admit their mistakes
@VK_ROXy true!