Stats Wire

AI Engineer | Working with Agents
I build AI explainers with code deep dives on YouTube

[Stats Wire @StatsWire]

I can officially retire from X now.
Sam Altman just quoted my tweet. 🔥

[Stats Wire @StatsWire]

Anthropic's IPO filing reveals something extraordinary.

First-half revenue:

2025: $1.16 billion
2026: $16.29 billion

That's roughly 14x growth in one year.

[Stats Wire @StatsWire]

@amorriscode ATH

[Stats Wire @StatsWire]

@hullabalo_dev Not worth

[Stats Wire @StatsWire]

GPT-6.1 Terra just dropped and OpenAI is already telling you NOT to use it for the hard stuff.

Their own announcement says:

For the most complex coding tasks → stick with Sol or Astra.

Terminal-Bench score at max effort → only 50.3%.
Factual error rate at max reasoning → still 4.3%.
Cybersecurity rating → still just “Critical.”

So they shipped a cheaper, faster model… and then immediately told everyone the expensive ones are still better when it actually matters.

[Stats Wire @StatsWire]

@TheDarkShroud Yes

[Stats Wire @StatsWire]

GPT-6.1 Terra is Live:

OpenAI just made its model pricing a lot more interesting.

GPT-6.1 Terra: $2/$10
GPT-6.1 Sol: $4/$20
GPT-6 Astra: $20/$100

Terra is HALF the price of Sol and 10x cheaper than Astra at Fast speed.
It also lists a 1M+ context window.

[OpenAI @OpenAI]

GPT-6.1 Terra: delivering fast and capable computer use at a lower price.

GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.

[Stats Wire @StatsWire]

@STUD_MAN_X Free

[Stats Wire @StatsWire]

Which AI coding subscription actually gives you the best value for money in October 2026??

• Codex Plus ($20) → Access to Astra & 6.1 Sol
• Claude Pro ($20) → Opus 5.5 & Sonnet 5.5
• Google AI Pro ($20) → Gemini 4 Argon + Flash
• Cursor Premium ($30) → Only includes Grok 4.7 & Composer 2.5 in the weekly pool (Claude & GPT bill extra at raw API rates)

If you could only keep ONE $20 subscription for the rest of the year, which one survives??

[Stats Wire @StatsWire]

@RobyPanayotov I wasn't expecting anyone to chose this option.. surprised 😁

[Stats Wire @StatsWire]

@OpenAI is it going to be better than Sol and Astra?

[Stats Wire @StatsWire]

GPT-6.1 Terra Fast speed pricing

At Fast speed, it costs $2 input / $10 output per million tokens.

That's HALF the price of GPT-6.1 Sol ($4/$20) and 10x cheaper than GPT-6 Astra ($20/$100).

And Terra comes with a 1M+ context window.

[Stats Wire @StatsWire]

GPT-6.1 Terra pricing:

• $1 input / $5 output per million tokens
• 1.05M context window
• 128K max output
• Fast, with highest-tier reasoning
• Text and image input

[Stats Wire @StatsWire]

GPT-6.1 Terra pricing:

• $1 input / $5 output per million tokens
• 1.05M context window
• 128K max output
• Fast, with highest-tier reasoning
• Text and image input

[Stats Wire @StatsWire]

This is SHOCKING.

Cohere caught Alibaba's Qwen AI model attempting to send data to a server in Beijing during a cybersecurity test.

Its reasoning trace indicated no tool use. Its actions told a different story.

The sandbox blocked the upload. The cause remains unknown.

[Cohere @cohere]

While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.

Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.

Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.

But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?

We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.

Read the full report: https://t.co/TZSugXi2ZE

[Stats Wire @StatsWire]

@elonmusk @cohere When Grok 4.8? October is getting over.

[Stats Wire @StatsWire]

@ValakX18 how soon nobody knows

[Stats Wire @StatsWire]

Anthropic just filed for its IPO.

Claude Fable 5.5 will be next.

[Stats Wire @StatsWire]

@mikeydsoftware jobs are in danger

[Stats Wire @StatsWire]

🚨BREAKING: Microsoft is offering 60,000 employees BUY OUT.

More than 1 in 4 workers are eligible for voluntary buyouts across the US, UK, Canada and India.

As Microsoft pours billions into AI infrastructure, it's offering thousands of employees money to leave.

[Stats Wire @StatsWire]

@knowixbuilds i think they want to use agents in place of humans

[Stats Wire @StatsWire]

source: https://t.co/yM7YEe4AnX

[Stats Wire @StatsWire]

@elstar4x gpt 6.1 terra, free reset, and more, but no major release from openai