It's the first time in the history of the world.
Silicon Valley: The Three KingdomsAIIt's the end of the war, the end of October!
Just today, there was an outcry on X that Anthropic was the ultimate crown jewel.Fable 5.5 is likely to come online tomorrow!
Almost at the same time, The Information went viral.The new model from Anthropic is already in the hands of the testers.。
Those who used it reveal that the next generation of Claude's mythical rise focused on the three great abilities of "thinking deeply, judging intuitively, and reading human intent".
And what's scariest is that it has developed a powerful "heart-reading" technique.Even if you don't say exactly what you want, it can get you what you really want.。
If you just let it fix one bug, it can solve a whole class of problems!
The Anthropic logo has not yet been officially unveiled, and neither OpenAI nor Google can sit down. Google's Gemini 4 Carbon is already running internally, and the new generation of OpenAI Terra is in the pipeline.
This October, the Battlefield of the Grand Model is doomed to smoke in flames.
In the next few weeks, the AI circle will be over again!
New Claude has learned to "read cardiology".
According to The Information, an early tester said that the reasoning and intuition behind this new model was that the new model was not a good idea.More than all the models currently on sale by Anthropic。
And the most disturbing thing about hair is that it's another real scratch.
Even if the information given by the user is small, even the goal is not clear, it is especially good at looking at the intentions and goals behind each other.
It's not hard to reason and intuitively understand, but "look through the goal you didn't mention".It's a direct link between humanity andAIThe depths of the soul of interaction。
It's a bug that creates a whole class.
For example, an insider who has worked on it says that you throw away a previous error and let it fix a bug.
The traditional model is a "headache headache", where a patch is crossed at the wrong place.
It's going to be a mess, and it's going to be tough.I'm going to try and solve a whole bunch of problems that have caused this bug.I'll go back and ask you not to repair it together.
Anyone who has written the code knows that the same problem is often hidden in several corners, repaired this, and reported elsewhere after two days.
Previously, AI had only modified one of your names, and the rest you have to look for yourself somewhere.The new model will look down on its own and present the same problem to you once and for all.。
And it's not self-evident.It's not going to change, it's going to change the scope, it's going to ask you a question, is it you or is it you?。
This is the kind of AI that you can give your life to.
For example, if you ask him how to fuck his ass, he doesn't play the game.I designed a set myself.AlgorithmsHow to predict the future of stock prices。
It's not surprising that test takers use the extremely human-colored word "intuition" to describe it.
It's been a long time coming, and it's been a long time coming, and it's been a long time coming, and it's been a long time coming, and it's been a long time coming, and it's been a long time coming.AI is officially entering the "innovator" stage.。
Writing a thermometer and catching a mining pigeon
This "thinking step" is real, and the actor Claude has already shown his hand.
In the past two days, a Mac user's experience on Reddit has been redirected to X, where a post has received 220,000 views.
He just asked Claude Code to help him write a little temperature monitoring program.
When Claude read the sensor data, he found that the readings were incredibly high and the CPU peaked at 112°C. It didn't cross the numbers to the menu bar, but instead turned around to see who was eating the CPU.
A mining program called xmrig is on the scene!CPUIt takes about 1,175 percent, which is equivalent to running all 12 cores.。
The program is hidden in /tmp/rigupdater, and the startup entry was buried on August 29, much earlier than its manual code.
It's not automatic, as it was when Bug was built.First, ask, "Can you clean up now?"I was just trying to see the temperature, and I found out the real killer of the computer cartoon.
And this is just an active Claude.
Fable 5.5, tomorrow's raid?Fable 5.5 winds have not stopped since the beginning of the month.
When a wave of developers broke out,Fable 5.1 was routed to a new, more powerful model.。
One of the most obvious signs is that it's even "replaced" Tibo who's answering the question, and this new, old model doesn't know it at all.
When they came, they were shocked and shocked.
Animation, 3D physical simulation, writing code from scratch, all of which can be breathed in.
Someone gave it just two suggestive words, adding only four lines, so that it could use JS to draw an animation that would have saved a person's life, even running for two hours and burning 27 million tokens.
It has been compared to the GPT-6.1 Sol frontal PK, which is also a 3D tug-of-war, with a much better stretch-rebound effect.
If the routing is just a gray test, the clue that emerged from last week's code library makes it feel as if the connection is in sight.
On October 13, a robot on GitHub registered a "claude-fable-5-5" under the name Microsoft Foundry.
On the same day, the open-source project LiteLLM added "azure_ai/claude-fable-5-5-5" to submit a statement that the name comes from the Azure model interface.
Although the post was later deleted, the clue was broken. Today, The Information reports, the fire has been reignited.
It is unclear whether the new model will be part of the Fable or Opus family. The former is the strongest of the Anthropic series, while the latter is the second.
However, the Executive Editor of The Information, Amir Efrati, came in person and asked on X: Is Anthropic about to release Fable 5.5?
This is what the executive director of his home said.The whole web bet is almost down, completely pressed on Fable 5.5.。
Some netizens have said that Anthropic has already suppressed OpenAI and that Fable 5.5 will be exclusive and unopposed.
It's not a blind guess that the netizens are shouting "tomorrow".Fable 5.1 and Opus 5.5 were both released on Tuesday.。
However, the reporter, Stephanie Palazzolo, has changed her mind, saying it will be in the coming weeks.
In other words, even if it doesn't come tomorrow, Fable 5.5 is only a few weeks away from release.
Google and OpenAI are at the same time glowing
Anthropic was a big hit, and the other two didn't do it either.
On Google's side, the Gemini 4 Argon was launched last week and the next Carbon card was running inside.
According to internal documents and screenshots seen by Business Insider, Carbon is on Google's internal programming platform, Jetski.
One employee said:It says "Feels like Opus 5.5".Another shouted: "Carbon is really good!"
The Information also gets the news that Carbon is one step ahead of Argon.
Employees brought the standard Opus 5.5. Programming is the most handy of the Anthropic discs.Google is going to be tough.。
Some developers have said that watching videos and testing whether web pages are running properly, while Gemini is still isolating itself.
The information reporter was also amused by Google's YouTube video training model, which probably helped a lot!
In terms of price, Google is just like Opus 5.5.Argon offers only half the price of Opus 5.5.After the offer expired on November 30, the game was released in the same format as Opus 5.5.
On the other hand, OpenAI is a new mid-range model.
According to sources, OpenAI may launch a new generation of Terra in the coming weeks. It is located between Sol and Luna, with more capabilities than the previous generation.The price is lower than the GPT-6.1 Sol.。
The GPT-6.1 Sol, which was launched at the end of last month and is close to the flagship Astra, is already a significantly lower price; the new Terra will be cheaper.
Interesting netizens have also discovered that Microsoft, a PR, has converted the GPT-5.6 Terra of 18 rating configurations into GPT-6.1 Terra.
OpenAI's next-generation flagship is also in preparation, with some circular Transformer architecture in the model. The entire company is stuck on it to make sure there are no major security issues when it goes online, so we'll have to wait a while.
The Three Kingdoms of Silicon Valley, Final Battle in October
Over the next few weeks, Fable 5.5, Gemini 4 Carbon, and the next generation of Terra will take to the stage.
But smart isn't depreciating everywhere.For example, in a long-term task involving millions of tokens, such as processing an entire code library,The top 10 models could more than double their traffic.In the field of programming, this figure is up to 72%.
The longer you live, the more complicated you are, the more valuable you are.This is the real battlefield of the Three Kingdoms.
Google adds that OpenAI is pricing, while Anthropic says it is "reading the mind". The hardest part of a long task is deciding where to go next.
When AI is able to find the real problem itself, people just have to say the goal and give it the rest.
By that time,AI has gone from being a tool for waiting for instructions to a partner for actively finding problems.。This is the key step towards AGI.
You haven't finished talking yet, it's already thinking about the next step.
RoboInd editorial channelResearch, robotics companies and market intelligence for cross-border sourcing.Explore the channel ↗
Share this
End of editorialExplore more from Global Robotics News ↗