We're planning a MindsHub subscription for everyday work with open-source AI models. Our goal is to cover a similar amount of non-coding work as paid ChatGPT and Claude subscriptions, at a lower monthly price. Your answers will help us decide what to include. About 2 minutes. The models, usage limits, and price are still being decided. docs.google.com/forms/d/e/1F…

It started as a weekend experiment. A month later, it's still adding to your AI bill. MindsHub helps you spot and cut wasted spend across a fleet of open and closed models. There's an always-free tier for your next idea. Try it: s.mdb.ai/4dWDzHF

Got a task your current AI model keeps getting wrong? Claude Opus 5.5, the GPT-6 family (Astra, Sol and Luna) and Grok 4.7 are available on MindsHub Inference. Bring that task and see how they handle it. Compare the results and costs to find the right fit for your app or agent. You can access them all through one API. Try them on MindsHub → (s.mdb.ai/4jnF05v)

Jev, TypeSafe’s decision model, joins MindsHub's free tier. API docs: docs.mindshub.ai/inference/d…

Introducing Free Tier for agents. mindshub.ai

Complete projects. Build software. Add intelligence to products. => mindshub.ai

This open source tool makes Claude Code and Codex work with open-source models. github.com/mindsdb/setfree

Manus is being pulled back into China’s orbit. Its users are dealing with the fallout. Before choosing another agent, ask a bigger question: can your entire AI workspace move with you? Our take: mindshub.ai/blog/manus-alter…

Are cheap LLMs actually cheaper? A five-times-lower token price often produces the same bill... We wrote a blog on how to compare models on cost per task, and measure your own. More 👉 mindshub.ai/blog/llm-cost-pe…

Interesting economic facts about AI: 1. Cumulative global spend on running models passed cumulative spend on training them (early 2026). 2. Prices for similar capability models fell over 99% between 2023 and 2026. For example, GPT 4 was $30 in 2023, Gemini Flash-lite is now $0.25. 3. Inference is now roughly 85% of enterprise AI budget and around two-thirds of global AI compute spend. 4. AI-first SaaS companies now spend 40–50% or more of revenue on inference and GPU, producing gross margins of 25–60% against the 75–85% that software valuations assume. WHat do you think - is AI just a bubble? 😉

Claude Opus 5 shipped today and took #1. But the top four models are now within 6% of each other. At that spread, price and speed decide, not rank. Updated comparison of 14 models: s.mdb.ai/4gWk9oU Opus 5 arrives in MindsHub Cowork on Monday.

We just changed a handle to @MindsHub

"Which AI model should I use?" changes its answer roughly weekly now. So we maintain a living comparison of the 14 that matter - what each is actually good at, what it really costs, and which ones you can run yourself. Updated today: s.mdb.ai/4waxD52

AI should be something you own, not just something you rent. Inspect it. Modify it. Fork it. Run it with your data, your tools, your way. That's why we built MindsHub Cowork — an open-source AI coworker you can actually make your own. Call it Claudio if you want ;) ⭐ s.mdb.ai/4vUMBfw 🌐 s.mdb.ai/4h0GktZ

Benchmarks won't tell you which LLM to use. They'll tell you which one wins a test that isn't your job. New guide: how to read the benchmarks, what actually matters (cost, speed, context, privacy), and when to self-host vs. use an API. s.mdb.ai/4420nkg

There's no "best" LLM in 2026 — only the best one for the job in front of you. So we built the plain-English comparison we wanted: GPT, Claude, Gemini, DeepSeek, Qwen, Kimi, Grok & GLM. What each is actually good at, and what it costs. s.mdb.ai/4vG3oCS