Ultraman podcast reveals shocking news: Astra's computer operation has reached human level

Author: Xinzhi Yuan

Today, exciting news has arrived.

In the latest podcast, Ultraman revealed that Astra's ability to operate computers has reached human levels!

Moreover, Fable 5.1 is not far away.

Some have discovered that Anthropic's latest model, Fable 5.1, has quietly appeared in the API routing of Amazon AWS Bedrock!

It is very likely that Astra and Fable 5.1 will be released around the same time.

The exciting moment is about to arrive.

Astra Operating Computers Has Reached Human Levels

Currently, this podcast has only released a little over a minute of a preview.

During the interview, the host posed the question: "Do you think AGI has arrived?"

Ultraman replied: "It's close, at least very close… AGI feels like a milestone, while superintelligence feels like a monster that can expand infinitely. As for Astra, it feels like it has completely reached human levels in operating computers."

In fact, the "Astra that has reached human levels" mentioned by Ultraman has already leaked a number of impressive demo tests recently.

Last Sunday, Astra revealed a series of stunning front-end demos.

For example, this test below.

With a simple prompt—"Build a magnificent sci-fi spaceship with an explorable interior"—without any multi-turn dialogue iterations, the AI directly "one-shot" created a complete 3D world.

As this influential figure said, AGI may not arrive in the form of "chatbots becoming smarter"; it may quietly appear in the form of "creating a complete world with a single sentence."

Moreover, OpenAI's ambitions go far beyond this.

Astra is not meant to compete with Fable in terms of who can chat smarter; it is born for "persistent agents."

The currently disclosed details of Astra are as follows.

Stunning 56K Tokens Inference: 3D Pagoda

Astra's latest output, a 3D voxel art piece, the pagoda, is truly stunning.

At maximum inference intensity, it consumed 56K inference tokens over 38 minutes, and Astra recreated a 3D building that achieved a very high standard in composition, color expression, and overall rendering effect.

This reflects extremely hardcore spatial imagination and code execution ability.

"Never Stopping" Digital Worker

It is no longer a passive tool that responds to your questions one by one. Astra can work continuously for days or even weeks; it has long-term memory and can remember your corrections.

Multi-Agent Collaboration and Cross-Tool Operations

It can spawn multiple agents to collaboratively tackle research-level mathematical proof challenges.

Even more frightening is that it can take over your computer desktop at "superhuman speed," collaborate across software, and even work alongside other AI agents and humans.

This confirms what Ultraman said in the podcast: "The ability of AI to use computers has officially reached human levels."

Fable 5.1 Appears in AWS Bedrock Routing

Moreover, today there is news that Fable 5.1 is not far away.

A developer has announced a significant discovery: Anthropic's latest model, Fable 5.1, has quietly appeared in the API routing of Amazon AWS Bedrock!

When developers called the model identifier for Fable 5.1 in the system, the system originally returned a "400" error (the provided model identifier was invalid, a made-up name); but now, it returns a "404" error (model not found).

This indicates that AWS's underlying system has officially recognized and accepted the identity of Fable 5.1, just waiting for the model itself to be uploaded.

This underlying change is completely consistent with the performance seen a few days before the release of Opus 5.1.

Additionally, over the past two weeks, many people have been gray-tested, and various signs indicate—

The release of Fable 5.1 has entered a countdown, likely within the next 3 to 4 days.

If Anthropic cannot seize the ecological position before Astra's release, Fable 5.1 may be eliminated.

Furthermore, some are testing Opus 5.1 in stealth mode, and the results are also quite good.

After Months of Testing, Why Does Fable 5 Dominate Over GPT 5.6?

Moreover, recently, the well-known community Bridgemind concluded after months of testing: GPT 5.6 is being dominated by Fable 5.

Over the past few months, he has been intensively mixing Fable 5 and GPT 5.6 Sol daily, and his experience is as follows.

GPT 5.6 Sol is a powerful backend machine but loses to "over-engineering."

In backend processing, GPT 5.6 Sol performs better. It is extremely fast, has lower invocation costs, and does better in data security.

Even in mathematical logic deduction, Sol has already significantly widened the gap with Fable and the previous generation Opus.

Once the work enters a deep and complex stage, Sol is a more powerful backend engine.

However, it has a fatal flaw; as long as the instructions are slightly loose, GPT 5.6 will suffer from the "over-engineering" occupational disease.

What should be a simple function, it insists on creating a huge, complex system for you, and you need to restrict it very precisely.

Meanwhile, Fable 5 remains the de facto No. 1.

It wins in two soul-piercing areas: front-end UI design and "one-shot" capability.

You throw a bunch of messy requirements at Fable 5, and it often delivers a "feels like it's done" product on the first attempt.

Its extremely high initial hit rate, combined with outstanding visual presentation and aesthetic taste, allows users to experience what true "smoothness" is.

This is why Fable 5 still firmly holds the top spot in developers' hearts—it's not only capable but also does it beautifully.

There are reports that Astra has now opened internal testing to some partners to collect feedback, and if all goes well, we will soon be able to experience this new species.

Next, who will take the lead, Astra or Fable 5.1?