LATEST DEVELOPMENT

Greetings, AI Explorers!

Today’s edition is packed with major AI breakthroughs and new developments. From “OpenAI to retiring GPT4.5” to “ask you for ID in future to access models”, the landscape is shifting fast.

As AI continues to integrate into everyday life, staying ahead of these developments is crucial. Let’s dive in!

Here’s what’s in store for you today:

• 🔚 OpenAI to Retire GPT-4.5 From API, Citing Cost and Efficiency

• ⚡ OpenAI’s New GPT-4.1 Family Focuses on Smarter, Faster Coding Agents

• 🦾 Hugging Face Acquires Pollen Robotics to Open Source a Humanoid Future

• 🪪 OpenAI to Require ID Verification for Access to Future Models

LATEST DEVELOPMENT

🔚 AI Model Lifecycle

OpenAI to Retire GPT-4.5 From API, Citing Cost and Efficiency

Just months after its release, OpenAI announced it will sunset GPT-4.5 from its API by July 14, nudging developers toward the newly launched GPT-4.1.

Key Points:

• GPT-4.5, OpenAI’s largest and priciest model to date, will no longer be accessible via API after July 14

• The model remains available inside ChatGPT’s research preview for paying users

• GPT-4.1, released today, is positioned as the API successor, delivering similar or better performance at lower cost

• GPT-4.5, codenamed Orion, used unprecedented compute and data but underperformed on some benchmarks

• Cost was a major factor: $75 per million input tokens and $150 per million output tokens — the most expensive model OpenAI has offered

• OpenAI said the decision helps them “prioritize building future models”

Why it matters:

OpenAI’s for-profit turn is looking to be a record-breaking one, and both company projections and investor wallets are signaling that the AI boom is not slowing down any time soon. The competition may be getting tougher, but as we’ve seen this week, no company dominates mindshare in the AI world quite like OpenAI.

⚡ AI Model Upgrades

OpenAI’s New GPT-4.1 Family Focuses on Smarter, Faster Coding Agents

OpenAI’s New GPT-4.1 Family Focuses on Smarter, Faster Coding Agents
OpenAI has launched a trio of new models, GPT-4.1, 4.1 mini, and 4.1 nano, that it says are tailor-made for software engineering and instruction following, with performance boosts and cheaper pricing.

Key Points

• GPT-4.1, 4.1 mini, and 4.1 nano are multimodal models designed for coding, with up to 1 million-token context windows

• All three are available via API only (not in ChatGPT)

• OpenAI says 4.1 excels at:

  • Frontend coding

  • Consistent tool usage

  • Reliable formatting

  • Reduced unnecessary edits

• 4.1 scored 52–54.6% on SWE-bench Verified, slightly behind Google’s Gemini 2.5 Pro (63.8%) and Anthropic’s Claude 3.7 Sonnet (62.3%)

• Pricing:

  • GPT-4.1: $2M input / $8M output

  • 4.1 mini: $0.40M / $1.60M

  • 4.1 nano: $0.10M / $0.40M

• GPT-4.1 scored 72% on a video comprehension benchmark (Video-MME, long videos without subtitles)

• Downsides: Model accuracy drops with larger input sizes (from 84% at 8K tokens to 50% at 1M tokens)

• OpenAI still aims for a long-term goal: building an “agentic software engineer” that can write, test, debug, and document entire apps.

Why It Matters
This release sharpens OpenAI’s coding chops just as competition from Google and Anthropic intensifies. GPT-4.1 might not top every benchmark, but its developer-oriented upgrades, massive context size, and cost efficiency hint at a future where AI truly becomes your junior (or senior) software engineer.

🪪 AI Policy & Access

OpenAI to Require ID Verification for Access to Future Models

In a move to tighten security and ensure safe use, OpenAI may soon require organizations to verify their identity before accessing its most advanced AI models via API.

Key Points

• Verified Organization is a new process that uses government-issued IDs

• One ID can verify one org every 90 days; not all orgs will qualify

• The goal: limit misuse, reinforce AI safety, and protect IP

• Comes amid reports of malicious usage, including alleged exfiltration by groups linked to North Korea and DeepSeek in China

• OpenAI blocked access from China last summer

• Verification applies to “certain future models,” not current offerings

Why It Matters
As AI models become more powerful, OpenAI is drawing a line: only trusted, verified orgs get access to the cutting edge. It’s a safety-first approach that also reflects growing concerns around espionage, model theft, and misuse at scale.

QUICK HITS

🛠️ Trending AI Tools

• 🌟 Firebase Studio – Build AI-powered SaaS tools fast with an all-in-one dev kit.

• 🎥 Pippit – Make videos, images, ads & avatars with AI—fast and easy.

• 🔎 Sherlock – AI proctor watches Zoom, Meet & Teams to stop cheating.

• 🧙‍♂️ Higgsfield – Create and edit stunning AI videos with cinematic control.* Sponsored Listing

💰 Big Money in AI

• 🚀 OpenAI’s Ilya Sutskever’s new AI startup hits $32B value

• 💼 Mira Murati’s AI startup eyes $2B seed round

• 🛠️ Incident.io raises $62M to help IT teams fix issues fast

📰 Everything else in AI today

• 🤖 ChatGPT Gets Long-Term Memory

• 🚨 OpenAI Retires GPT-4

• 🚗 Google Maps Gets AI Boost

• 🌐 DeepMind Merges Gemini & Veo

• 🔎 Deep Research Lands on Gemini 2.5 Pro