GLM 4.7 Outperforms GPT-5.2

Welcome to another edition of AI PlanetX.

GLM 4.7 tops GPT 5.2 and Claude 4.5 Sonnet in global AI benchmarks; AI helps uncover cancer survival links, while DeepMind CEO challenges skepticism on general intelligence.

• Hottest AI News

• Top AI & SaaS Tools

• AI Tutorial: Turn Any Webpage into Post Using Comet

• Top AI & Tech News

• AI Art Spotlight

• Prompt of the Day: GPT-5.2 Prompting Guide by OpenAI

• AI Video Tutorial

• Course of the Day: Governing AI Agents

Hottest AI News

Zhipu

GLM 4.7 Outperforms GPT 5.2 and Claude 4.5 Sonnet in Global Benchmarks

Zhipu AI (Z AI) has launched GLM-4.7, a flagship open-source SOTA model that rivals—and in some benchmarks surpasses—top models like GPT-5.2 and Claude 4.5 Sonnet in coding and complex reasoning.

Details:

• This release ranks #1 among open-source models on LMArena Code Arena. It also beats Claude 4.5 Sonnet on LiveCodeBench V6 (84.8) and surpasses GPT-5.1 and GPT-5.2 on AIME 2025, τ²-Bench, and HLE reasoning, with a 38% gain

• Technically, the model supports a 200K context window with up to 128K output, runs at 55+ tokens/sec, and includes a Deep Thinking mode plus optimizations for agentic coding in tools like Claude Code, Cline

• Developers can use the global Open Platform via Python and Java SDKs, with 4096-token requests, adjustable parameters like temperature, and both standard and streaming APIs

By releasing GLM-4.7 with strong reasoning and conversational depth, Z AI aims to meet growing demand for powerful, easy-to-deploy language models.

Run IRL ads as easily as PPC

AdQuick unlocks the benefits of Out Of Home (OOH) advertising in a way no one else has. Approaching the problem with eyes to performance, created for marketers with the engineering excellence you’ve come to expect for the internet.

Marketers agree OOH is one of the best ways for building brand awareness, reaching new customers, and reinforcing your brand message. It’s just been difficult to scale. But with AdQuick, you can plan, deploy and measure campaigns as easily as digital ads, making them a no-brainer to add to your team’s toolbox.

You can learn more at AdQuick.com

AI in Healthcare

RNACOREX Finds Clear miRNA and mRNA Links in Cancer Survival with AI Support

Researchers at the University of Navarra developed RNACOREX, an open-source tool that maps interpretable miRNA–mRNA networks linked to cancer survival. Tested on 13 tumor types, it combines curated databases and gene expression to produce explainable molecular maps.

Details:

• It ranks and filters miRNA–mRNA interactions using curated databases plus TCGA expression data to build probabilistic regulatory networks and reduce false signals

• Tested on 13 cancers (e.g., breast, colon, lung, stomach, melanoma, head & neck), it predicts patient survival comparably to advanced AI while revealing the molecular interactions behind predictions and highlighting shared patterns and candidate targets

• RNACOREX is open-source (GitHub, PyPI), provides automated database tools, received funding from Gobierno de Navarra and ERA PerMed, and will expand with pathway analysis and extra molecular layers

RNACOREX offers an explainable, reachable alternative to black‑box models, helping accelerate precision cancer research and target prioritization.

Top AI & SaaS Tools

• Brilliant Directories (Life-time Deal): AI-powered CMS and CRM that simplifies building, launching, and monetizing membership and directory websites without coding or plugins

• Surgeflow: Browser-based automation tool that transforms multi‑tab browsing into organized, transparent workflows [F-R-E-E]

• Flowdrop V1.1: Turns text into no-code automations that connect your apps and add human approval, real‑time checks, and integrations like Gmail, Slack, and Sheets [F-R-E-E]

• ConnectMachine: AI contact manager that creates shareable digital business cards, auto-organizes connections, and lets you query your network via voice or text [F-R-E-E]

• Incredible: An always‑on AI agent that automates daily tasks with near‑zero hallucinations, large memory, and hundreds of integrations [F-R-E-E]

A Framework for Smarter Voice AI Decisions

Deploying Voice AI doesn’t have to rely on guesswork.

This guide introduces the BELL Framework — a structured approach used by enterprises to reduce risk, validate logic, optimize latency, and ensure reliable performance across every call flow.

Learn how a lifecycle approach helps teams deploy faster, improve accuracy, and maintain predictable operations at scale.

Read the Full Guide

AI Tutorial

How to Turn Any Webpage into a Post Using Comet

Managing multiple social media channels can feel overwhelming — but with Comet, you can turn any webpage, video, or thread into shareable content without ever leaving your browser.

1. Download Perplexity Comet

   Install Perplexity Comet as a standalone app for Mac or Windows from the official Perplexity website. Once installed, launch it just like any other browser.

2. Set up workspace

   Sign in using your Perplexity or Google account. Then, customize your workspace — choose whether you want Comet to prioritize short summaries, social media captions, or detailed write-ups.

3. Visit content source

   Go to a Twitter/X thread, LinkedIn post, Reddit discussion, or even a YouTube video. Comet automatically detects the main topic of the page you’re viewing.

4. Tap Assistant icon

   You’ll see a Comet Assistant button at the top right corner of your browser. Press it to activate Comet’s contextual understanding of the page.

5. Type clear prompt

   Ask Comet to perform a task related to the page.

   Example prompts:

   • “Turn this thread into a LinkedIn post.”

   • “Summarize this YouTube video in 5 bullet points.”

   • “Explain the main idea of this article like I’m 15.”

   • “Write a tweet thread based on this blog post.”

6. Refine or expand AI’s response

   Once Comet generates a summary or draft, you can ask it to:

   • Add more detail to specific sections

   • Rephrase for a casual or professional tone

   • Extract quotes or timestamps from videos

7. Review before posting

   Check the content Comet creates — summaries, captions, or posts — for tone, accuracy, and style before publishing them to your connected accounts.

8. Publish directly or copy

   Comet can post directly to your Twitter, LinkedIn, or Facebook account. Or, you can copy the generated text to reuse elsewhere, like in newsletters or captions.

Tip: Try using different prompt styles — sometimes phrasing it as a question (“What’s the main takeaway here?”) gives better results than a command.

Top AI & Tech News

• Demis Hassabis (DeepMind CEO) rejects Yann LeCun’s claim that “general intelligence is an illusion”

• Z-Image Turbo become the top text-to-image model, overtaking FLUX.2 [dev], HunyuanImage 3.0, and Qwen-Image in the Artificial Analysis Image Arena

• OpenAI’s compute margin — the share of revenue remaining after the costs of running AI models for paying users — rose sharply to about 70% in October

• Nvidia told Chinese clients it plans to start shipping its H200 AI chips to China before the Lunar Year (mid‑February), using existing stock to fulfill initial orders

AI Art Spotlight

Create Your Own Photo Like This Using the Prompt Below:

A playful yet photorealistic 3D render of the Pinterest logo, made from soft foam material. Rounded edges, slight compression marks, and smooth matte surface. Bright cheerful colors, minimal studio background, soft diffused lighting, product photography style.

Model: Nano Banana Pro

Prompt of the Day

GPT-5.2 Prompting Guide by OpenAI

This guide shows how to migrate prompts to GPT-5.2 for production agents, with practical patterns for controlling verbosity, enforcing scope, handling long contexts, and reducing hallucinations. It covers GPT-5.2's key improvements—better instruction adherence, tool grounding, and structured reasoning—plus concrete examples (verbosity specs, tool rules, extraction schemas) and migration steps to preserve behavior when upgrading.

Top AI Video Tutorial

Create Seamless AI Films from a Single Prompt — Consistent Characters and Scenes

AI Course of the Day

Governing AI Agents

This course shows how to integrate governance across an agent’s lifecycle—from permission control and policy enforcement to runtime monitoring and observability—using a practical Databricks dataset example. You’ll learn to build agents that manage data responsibly while maintaining visibility and safety.