the launch mentions cybersecurity false positives dropping 60%, but doesn't say whether that's from genuinely better judgment or a loosened threshold - a published tradeoff number (false negatives, not just false positives) would make that claim easier to trust rather than take on faith.
Netlify
Hey PH community 👋
Excited to hunt Claude Fable 5.1 today!
Anthropic has shipped a lot of models, but this one reads like a direct response to what builders kept asking for: better results on long, agentic work, real answers on price and data retention, and safeguards that get out of the way for legitimate use.
Here's a terrific example of what it can do. Alex gave it a picture of a property lot. It designed a house for the lot, rendered it, and produced a cinematic walkthrough 🤯
The pricing move is what stands out to me.
Cache reads got cheaper, so typical workloads cost about 25% less than Fable 5, and highly agentic runs can save up to 45%. Pair that with Low and Medium effort modes that match or beat Fable 5 output, and you're getting more model for less money.
Cybersecurity false positives are down 60% too, which anyone who has hit a wall mid-task will appreciate.
It's also live day one across the major AI gateways, including Netlify's, so you can drop it into an existing stack without touching your keys or config.
Go take it for a spin and tell the team what you build in comments 🚀
Mastra
i'm seeing things about fable i couldn't believe
Netlify
@fmerian yep same
So many demos on X
DiffSense
@fmerian Love it. Using Fable 5.1 max right now in Claude design....wait...used up my session quota....okay pingpong for a couple of hours to wait for it to reset. I will buy another subscription tmrw I guess. How many subscriptions do I need? is 2 enough? 3? :D
DiffSense
@fmerian LOL. It was so good. I got another subscription just now. Don't tell Anthropic. I'm not sure its "allowed" 😅
I like seeing more attention on scientific and research workflows. That’s an area where stronger models could be genuinely useful.
OpenObserve
@thisiskp_ Love Fable, the example is really terrific !
Does the model perform better when working through long research tasks with multiple sources?
LLMagnet
The pricing change is the interesting part. Better models are great until you actually run agents all day and see the bill 😅 Curious how well Medium effort holds up on long coding tasks vs fable 5
The knowledge-work angle caught my attention. Handling analysis, research, and coding in the same workflow could be quite useful.