GPT-5 Drops, GPT-OSS Goes Public, and n8n’s Pricing Sparks Debate

Two big updates. One wild week.
AI just threw us a plot twist, and automation fans are feeling the whiplash.
The Double Whammy Nobody Saw Coming
This week wasn’t just another round of “AI is taking over the world” headlines. Nope. We got hit with two game-changing, debate-sparking moves; one from OpenAI, the other from n8n.
OpenAI dropped the shocker of the year: open-weight GPT-OSS models and the long-awaited, “PhD-level” GPT-5.
n8n—beloved by automation nerds, rolled out a new pricing model that’s got self-hosted users… let’s say… audibly sighing.
Whether you’re tinkering with personal projects, scaling a startup, or just keeping tabs on the AI arms race this week’s developments matter. Big time.
OpenAI’s Bold Moves: Freedom Meets Firepower
1️⃣ GPT-OSS: Open-Weight is Back (For Real This Time)
Remember when open-weight models felt like a relic from the GPT-2 days? Well, dust off that nostalgia—OpenAI is giving us GPT-OSS, a suite of models you can actually run, tweak, and fine-tune yourself.
License: Apache 2.0 (yes, you can build your own on top of it).
Privacy: Keep your data local—no cloud dependency.
Accessibility: The smaller
gpt-oss-20b
runs on a decent laptop with 16GB RAM.
For years, large language models were locked behind API paywalls. Now? We can peek under the hood, make adjustments, and build hyper-specialized AI without begging a corporate endpoint for permission.
The dev community? Ecstatic. This feels like OpenAI extending an olive branch to open-source lovers—while still keeping the crown jewels (like GPT-5) proprietary.
2️⃣ GPT-5: The Apex Predator of AI
Launched: August 7, 2025.
Tagline: “PhD-level expert.”
Reality: Scary good at reasoning, coding, and not hallucinating (claims of 80% fewer fabrications).
It’s multimodal, seamlessly handling text, images, and even voice. It can code entire UIs from a vague prompt and navigate complex logic with ease.
But here’s the twist:
Experts: “This is a massive leap.”
Some users: “Feels… less fun to talk to?”
Reddit: A mix of awe and “bring back the old personality!”
It’s the eternal AI trade-off: brains vs. charm.
n8n’s Pricing Plot Twist: From Steps to Executions
1️⃣ The New Model
n8n shifted from limits on steps and users to a per-workflow execution model.
Execution = One complete run of your workflow, no matter how many steps it has.
Community Edition (self-hosted) = Still free.
Paid Plans: Business & Enterprise now charge based on executions, even for self-hosted setups.
2️⃣ The Backlash
For cloud users? Predictable billing might be nice.
For self-hosted power users? Not so much.
The criticism:
“We already pay for our servers, power, and maintenance—why should we pay again per execution?”
High-volume automators are doing the math and finding six-figure annual bills for heavy workloads. Some see it as punishing success, others as a nudge toward vendor lock-in.
Community forums? Let’s just say… lively.
What This Means for the Future
🆓 Open-Weight’s New Chapter
GPT-OSS could spark a wave of indie AI tools, local deployments, and niche fine-tuned assistants. The challenge: balancing freedom with safeguards against misuse.
🏔 GPT-5 and the Road to AGI
We’re not at true AGI yet GPT-5 can’t continuously learn post-release but it’s a major step. Expect it to disrupt industries, automate complex work, and, yes, reshape job markets.
💰 n8n’s Balancing Act
If n8n listens to its self-hosted community, it might introduce hybrid pricing mixing execution limits with feature-based tiers. If not? We may see an exodus to other open automation tools.
The Takeaway
In just one week, the AI landscape shifted in two very different directions:
OpenAI loosened the gates (a little) with GPT-OSS while flexing with GPT-5.
n8n tightened theirs, betting that execution-based billing will scale.
One feels like more control in your hands.
The other feels like more control in theirs.
Either way, the only constant here is change—and if you’re building in this space, adaptability isn’t optional.
Subscribe to my newsletter
Read articles from Mark P directly inside your inbox. Subscribe to the newsletter, and don't miss out.
Written by
