In partnership with
🖱
OpenAI just made your life easier.
GPT-Image-2 now generates images with transparent backgrounds natively no more post-processing headaches. Just flip a parameter and you're done.
But wait, there's more:
Anthropic dropped 355 free resources in its new Claude Academy
DeepSeek's experimental V4 Flash model can now see images
I'm Alex. Welcome to L8R by Innov8.
Let's dive deep 🐰
In today's post:
OpenAI's GPT-Image-2 now generates transparent backgrounds natively
Anthropic drops 355 free resources with Claude Academy launch
DeepSeek's V4-Flash goes multimodal with experimental vision model
GPT-Image-2 Finally Gets Native Transparent Backgrounds
You can finally stop using clunky background removal tools for your AI images.
OpenAI just gave GPT-Image-2 native support for transparent backgrounds through its API.
This means you get perfect cutouts for your designs right from the start.
The details:
Developers just need to add the background="transparent" parameter to their API requests.
The feature creates PNG or WebP files with a true alpha channel.
It perfectly captures tricky details like glass, hair, and translucent edges that normal tools ruin.
Watch out for your prompts, because asking for a specific scene or backdrop will override the transparency setting.
Why it matters:
Manual background removal is a massive time sink for designers and developers. This update kills that extra step. You can now instantly drop high-quality AI assets into websites, slide decks, and e-commerce stores. It fixes a huge missing piece that users have complained about since GPT-Image-2 launched in April 2026.
💡 L8R's Take:
It is crazy that it took OpenAI this long to add a basic feature like transparency. Still, native alpha channels mean AI image generation is finally ready for real work.
From Our Partner
Your team’s new support player
Stax Payments, a Hubspot customer, now books 80% more meetings with the exact same team they had before.
No new hires. No new territories. Just HubSpot's Prospecting Agent finding the right accounts, surfacing the right contacts, and drafting outreach that actually gets responses.
Your best rep is great because they know who to go after and when. Prospecting Agent gives every rep that same starting point.
Anthropic Just Dropped 355 Free AI Courses
Learning how to talk to AI just got a massive upgrade.
Anthropic launched a huge new learning hub called Claude Academy, and it is totally free.
They want to teach you how to use AI like a true pro, instead of just memorizing cheap prompt tricks.
The details:
Anthropic opened the academy with 355 free resources, including videos and hands-on exercises.
The courses cover everything from basic daily tasks to building complex agents with Claude Code and the Model Context Protocol.
Anthropic is teaching the exact same '4D AI Fluency Framework' they use to train their own internal employees.
You can earn digital certificates for finishing courses and install a special Claude Skill to get personalized learning paths.
Why it matters:
Prompting used to be about weird hacks and secret words. Now, Anthropic is shifting the focus to real AI fluency and good judgment. They actually want you to know when not to use AI. This massive free library makes it much easier for regular people and big companies to get real value out of Claude.
💡 L8R's Take: This is a huge win for all of us.Go straight to the source, learn Anthropic's internal framework for free, and get yourself certified.
From Our Partners
100 real-world tasks. 8 tool categories. 1 complete DevOps stack.
Master Git, Docker, Kubernetes, Linux, CI/CD, Terraform, and monitoring through structured tasks across real-job scenarios, and earn a shareable credential that proves you can apply what you’ve learned.
No lectures, no theory. Build a public portfolio as you go. Earn a verified badge. Free.
DeepSeek Gives Eyes to Its Flash Model
DeepSeek just gave its lightning-fast V4-Flash model the ability to see.
The new experimental vision model processes images and screenshots without losing its top-tier reasoning skills.
It is a direct shot at Anthropic's massive Opus 4.8 model but at a fraction of the cost.
The details:
The new DeepSeek-V4-Flash-Vision-Exp model uses 284 billion total parameters but only activates 13 billion during use to stay fast and cheap.
DeepSeek claims this new model approaches the performance of Anthropic's Opus 4.8 in multimodal agent tasks.
In internal tests, the model actually beat Opus 4.8 on three major benchmarks, including DeepSWE and Agents' Last Exam.
You can feed it mixed text and images, and each image costs a maximum of 384 tokens at standard V4-Flash pricing.
The model went live on the DeepSeek API platform on August 21, 2026.
Why it matters:
Vision models are usually expensive and slow. DeepSeek is changing the game by adding image support to its high-speed Flash tier. This lets you build smart AI agents that can look at screens and make decisions without burning through your cash. It proves you do not need giant, expensive U.S. models to get top-tier multimodal performance.
🚀 Quick L8R Summary
GPT-Image-2: OpenAI's image model now generates transparent backgrounds natively no more post-processing hacks.
Claude Academy: Anthropic dropped 355 free learning resources to turn you into a Claude power user.
DeepSeek V4-Flash-Vision: DeepSeek's experimental multimodal model now reads images while keeping its 13B-active architecture lean.
📩 Innathe L8R engane undarunnu 👇?
We read every reply - just reply to this email and let us know how we can improve !
Appo adutha L8R il kanaam bie…👋