Instagram | Sponsor this newsletter | Let’s Connect!

My fellow AI explorers

GPT-5 has landed, Claude leveled up, Gemini flipped “Deep Think,” and even video editing got a Photoshop-for-video moment. I’ve been hammering on GPT-5 all week and, yes, it’s the real deal—faster, sharper, and noticeably better at knowing when to think hard vs. when to just answer. Let’s dive.

In today’s edition:

• 🤖 GPT-5’s unified “think or go fast” model, smarter routing, and safer answers

• 🧠 The counterpunch from Claude Opus 4.1 and Gemini 2.5 Deep Think

• 🎬 Runway’s Aleph turns text prompts into video VFX

Must See AI Tools

• 💰 Payman: AI That Pays Humans. Over 10,000+ signed up for the beta

• 💫 SubMagic: An AI tool that edits short-form content for you! (Get 10% off using code “uncoverai” at checkout)

• 🎤 11Labs: #1 AI voice generator (Click Here to get 10,000 free credits upon signing up!)

• 🤖 ManyChat: Automate your responses & conversations on IG, FB and more! (Click Here to get first month for free)

• 🎙️ Syllaby: The only social media marketing tool you’ll ever need - powered by AI! (Get 25% off the first month or any annual plan with code “UNCOVER” at checkout)

Open AI

GPT-5 is here: unified, smarter, and available to everyone

One model, two gears. GPT-5 ships as a unified system that decides in real time whether to respond instantly or invoke deeper reasoning—plus a manual “think longer” control when you want it. OpenAI also introduced a smarter router that learns from real usage to pick the right path (and tools) per task. The result: higher accuracy, better planning, and far fewer “uhhh actually…” moments.

What’s new and useful:

• Hybrid brain: quick answers for easy stuff, deeper chains for hard problems; you can force “think hard” when needed. See the routing details in the launch blog above.

• Big factuality bump in day-to-day queries and noticeably better honesty about what the model can’t do. More in OpenAI’s post on how they’re optimizing ChatGPT.

• Coding power: strong performance on real-world repos and function calling; early third-party testing aligns with this.

• Safety shift: from binary refusals to “safe-completions” that stay helpful while respecting guardrails (explained in the launch and optimization posts).

• Rollout and access: GPT-5 is live in ChatGPT (free users included, with higher limits on Plus/Pro) and is already surfacing in partner stacks like Microsoft and GitHub over time.

Why it matters
For most workflows, GPT-5 feels like two teammates in one: a fast respondent and a careful reasoner, orchestrated by a router that increasingly “gets” your intent. In practice that means cleaner drafts, better plans, fewer wild guesses, and more reliable tool use. Early enterprise tests echo this—Box reports a solid bump on document understanding and extraction with GPT-5 live inside Box AI Studio. Expect a wave of “it finally works” moments across document AI, analytics, and sprawling codebases.

Prediction
Agents get way more useful, fast. With routing + safe-completions + stronger code edits, multi-step “research → plan → do → verify” loops graduate from demos to dependable daily tools.

Founders…

🤖 Need an AI Agency to Help Your Business Implement AI Solutions?

Align Coach has an expert team that understands how to help companies identify how AI and automation can transform their business. Get 20%+ more productivity and results from your team now.

Click here to learn more and get your company’s strategic plan.

Anthropic

Claude Opus 4.1 and Gemini 2.5 Deep Think: the counterpunch

While GPT-5 grabbed the spotlight, Anthropic and Google shipped meaningful upgrades.

• Claude Opus 4.1: better coding, reasoning, and autonomous task handling—Anthropic recommends upgrading from 4.0. If you live in Claude Code, you’ll feel the repo-wide search/analysis boost.

• Gemini 2.5 Deep Think: Google’s “parallel thinking” adds longer reasoning and new RL techniques; the advanced track hit gold-medal-level math in the IMO demo. Stronger step-by-step problem solving inside Gemini.

Takeaway
We’re in a “thinking model” arms race. GPT-5 currently sets the pace for accuracy + coding + router UX; Claude still shines in agentic, long-running multi-task jobs; Gemini’s Deep Think narrows gaps on math and structured reasoning. Pick per job, not per brand.

AI SaaS Founders

🚨Want Millions of Impressions For Your AI SaaS, Done For You?

At uncovernews.co, we specialize in getting AI SaaS products the attention they deserve through strategic influencer marketing campaigns designed to drive millons of impressions at the fraction of the cost!

Get Your AI Startup’s News or Product In Front of Millions Quickly

AI Video

Runway Aleph: Photoshop energy for video

Runway’s new Aleph model lets you add/remove/retarget objects, restyle scenes, and even change lighting and angles from plain text—on your existing footage. It’s multi-task, context-aware, and already available to paid users.

Think “make it rain,” “set the mic on fire,” or “swap the background to neon Tokyo,” without timeline wrestling. See examples in the Runway teaser.

Why you care
For content teams, Aleph turns weeks of VFX iteration into minutes of prompt-driven shots. Expect social teams and indie creators to push wilder visuals without touching heavy compositing suites.

30-Second AI Play

Build a phased refactor plan for a messy repo with GPT-5

1. Zip your repo (or share a link) and paste a short brief: current state, constraints, and your “must-keeps.”

2. Prompt GPT-5 Pro: “Analyze this codebase. Propose a 3-phase plan to harden module management, minimize risky rewrites, and ship phase 1 in two weeks. List risks and validation tests per phase.”

3. Clamp scope: “Constrain phase 1 to what two engineers can finish in 10 days. No greenfield rewrites.”

4. Ask for diffs and scaffolds: “Generate file-by-file edits and migration scripts for phase 1 only.”

5. Verify: “Create a test checklist and success metrics. What breaks if we skip step X?”

Why this works
GPT-5’s routing + reasoning is unusually good at meeting you where your project is—shipping a realistic, near-term plan instead of a 6-month moonshot.

Other Relevant AI News!

OpenAI released widely loved open-weights—check out the open models drop. Great math/efficiency trade-offs, with early cloud integrations following.

Gemini’s cute side project, Storybooks, auto-creates illustrated children’s books (cover, pages, art, narration) from a single prompt.

DeepMind’s Genie 3 turns prompts into interactive, navigable worlds—early, but a big swing for embodied agents and XR.

Kaggle launched Game Arena and a public chess benchmark. In an exhibition, o3 bested Grok 4—fun signal on strategic reasoning.

ElevenLabs rolled out shockingly clean vocal tracks with ElevenMusic. Expect licensing clarity to pull more creators in.

Anthropic announced upgrades you’ll feel in day-to-day coding and research—read the Opus 4.1 post.

Golden Nuggets

• 🚀 GPT-5’s router + safe-completions make agents feel less “demo” and more dependable.

• 🧩 Claude 4.1 and Gemini Deep Think keep pressure on deep reasoning—use the best model per workflow, not a single hammer.

• 🎥 Runway Aleph shrinks VFX from days to prompts—expect a flood of high-concept video from small teams.

Until our next AI rendezvous,

Anthony | Founder of Uncover AI