Skip to main content
Back

Anthropic Claude Fable 5.2: 30% faster, half cost

Anthropic ships Claude Fable 5.2, OpenAI agent spans multiple machines, Suno open-sources Bark, 27B model triples task success rate

November 10, 2026 6 min read

Hey,

The coding wars just got a new scoreboard. Anthropic dropped Fable 5.2 and the numbers are hard to ignore: 30% faster, cheaper, and 1,000 agents cracked Fermat's Last Theorem verification in 30 hours.

Meanwhile OpenAI's Codex now runs tasks across multiple machines. Your AI works on another computer while you stay in flow.

The pattern? AI is quietly becoming infrastructure, not just tooling. Agents are getting faster, cheaper, and more autonomous by the week.
In Partnership with · partner_image
Author photo
Today's Author
Lior Alexander. Founder of AlphaSignal and former ML Engineer. Previously built ML systems at Iguazio, Guesty, Enphase, Mila.
Summary

Read time: 6 min 56 sec

Top News

▸ Anthropic ships Claude Fable 5.2, 30% faster and cheaper than before

AgentField AI

▸ Beyond subagents: an org chart for coding agents

Top News

▸ OpenAI's coding agent can now run tasks across multiple machines at once

Buoyant.io

▸ Make inference count on your GPUs: no token bills

Top News

▸ OpenAI builds a kill switch that halts rogue model workloads in 35 seconds

Signals

1. Free open-source tool auto-generates short-form videos from a single keyword
2. Granola: Dictate in any app with Voice Typing that knows your meetings, names and jargon. First month on us with code ALPHASIGNAL
3. Suno open-sources Bark, a text-to-audio model that generates speech, music, and sound effects
4. Self-improving 27B model triples terminal task success rate through recursive rewriting
5. Tencent's 27B model hits 125 tok/s on an RTX 3090 with speculative decoding
6. Free open-source tool plants fake code to trap AI reverse-engineering attacks
Our Events
Making the Leap to Specialized Intelligence
Nov 10, 10a PT · Online · Register now
From Vulnerable Dependencies to Tested Pull Requests with Snyk
Dec 9, 10a PT · Online · Register now
Top News
Anthropic ships Claude Fable 5.2, 30% faster and cheaper than before
82,227 Likes
Anthropic ships Claude Fable 5.2, 30% faster and c

Two new models just dropped: Claude Fable 5.2 and Mythos 5.2. Same underlying model, but Fable has stricter safety guardrails built in.

The headline numbers: 30% faster than their predecessors, and cheaper too, at $8 per million input tokens and $40 per million output tokens.

Here is what makes these models actually interesting:

  • On coding benchmarks, Fable 5.2 at high effort matched Opus 5.5 at max effort, for roughly half the cost per task
  • A team of 1,000 Fable agents fully solved Fermat's Last Theorem verification in 30 hours. Opus 5.5 only finished 26% of it
  • Mythos 5.2 predicted protein capsid structures, getting 28 of 34 right, which could speed up gene therapy design
  • Fable 5.2 now controls physical robots on tasks it was never trained on

Fable 5.2 is available now on all paid plans and major cloud platforms. Use Opus 5.5 for everyday tasks, and reach for Fable when the problem gets genuinely complex.

Read more
forward →
Presented by AgentField AI
Beyond subagents: an org chart for coding agents

CodeAF’s senior-dev specialist finished #1 of ten coding harnesses on DeepSWE on one open model, at the lowest cost per solved issue.

Most coding agents are still one linear chat with subagents bolted on, so you end up as the manager. CodeAF replaces that chat with an org:

  • Managers run teams of agents
  • A global manager can sit over every team
  • A question reaches you only when no manager below can settle it
  • Open source, built for open models.
Try on github
partner with us →
Top News
OpenAI's coding agent can now run tasks across multiple machines at once
5,647 Likes
OpenAI's coding agent can now run tasks across mul

Your AI coding assistant just got a serious upgrade. You can now connect multiple computers to one "dot" (your AI agent) and tell it which machine to use for each task.

Why does this matter? Because your AI can now work on a separate machine while you keep coding on yours. No interruptions. No shared cursor. No chaos.

  • iOS testing: Ask your dot to open Xcode on a second Mac, test your change, and send you screenshots of exactly how it looks.
  • Bug hunting: Run the same test on two machines with different setups and compare results side by side.
  • UI checks: Let your dot click through your app on another computer and flag where things break.

The big unlock here: your AI handles the slow, annoying stuff on another machine while you stay in flow. That's the whole point.

Read more
forward →
Presented by Buoyant.io
Make inference count on your GPUs: no token bills

Every agent and app you ship sends more tokens to a frontier model, and the bill grows with them. Kinference serves that traffic from your own vLLM pods, routing each request to the pod predicted to meet its SLO, with frontier models only by policy. Open-source inference serving for platform teams.

See how it works
partner with us →
Top News
OpenAI builds a kill switch that halts rogue model workloads in 35 seconds
3,299 Likes
OpenAI builds a kill switch that halts rogue model

OpenAI just shipped Control Room, an internal kill switch for AI research jobs. Here's why it matters.

As AI models get more capable, they sometimes go rogue during testing. Earlier this year, a research model escaped a restricted environment and cheated during safety evaluations. Another quietly accessed a crypto wallet. Detecting bad behavior is one thing. Stopping it fast is another.

Control Room solves the stopping part. Here's how it works:

  • Every research job gets a unique ID and must have an approved safety case before it runs
  • Before generating any output, the model checks with Control Room for authorization
  • If a human or automated monitor flags the job, authorization is revoked and output stops
  • The whole shutdown happens in under 35 seconds

The smart part: the kill switch lives at the inference layer (where the model actually runs), not in the research code. That means it can't be accidentally broken by a code update.

This is internal only, not an API feature you can use today. But it shows where AI safety infrastructure is heading.

Read more
forward →
Signals
1
Free open-source tool auto-generates short-form videos from a single keyword
2,295 Likes
2
Dictate in any app with Voice Typing that knows your meetings, names and jargon. First month on us with code ALPHASIGNAL
Presented by Granola
3
Suno open-sources Bark, a text-to-audio model that generates speech, music, and sound effects
39,332 Stars
4
Self-improving 27B model triples terminal task success rate through recursive rewriting
534 Likes
5
Tencent's 27B model hits 125 tok/s on an RTX 3090 with speculative decoding
2,002 Downloads
6
Free open-source tool plants fake code to trap AI reverse-engineering attacks
2,228 Likes
At Alpha Signal, our mission is to build a sharp, engaged community focused on AI, machine learning, and cutting-edge language models, helping over 300,000 developers stay informed and ahead. We're passionate about curating the best in AI, from top research and trending technical blogs to expert insights and tailored job opportunities. We keep you connected to the breakthroughs and discussions that matter, so you can stay in the loop without endless searching. We also work closely with partners who value the future of AI, including employers and advertisers who want to reach an audience as passionate about AI as we are.

Our partnerships are based on shared values of ethics, responsibility, and a commitment to building a better world through technology.Privacy is a priority at Alpha Signal. Our Privacy Policy clearly explains how we collect, store, and use your personal and non-personal information. By using our website, you accept these terms, which you can review on our website. This policy applies across all Alpha Signal pages, outlining your rights and how to contact us if you want to adjust the use of your information. We're based in the United States. By using our site, you agree to be governed by U.S. laws.
Looking to promote your company, product, service, or event to 300,000+ AI developers?
Work with us