StepFun's flagship model is free to use right now.

2026-10-14 03:51

StepFun's flagship model is free to use right now. WorkBuddy just plugged in Step 5 Preview, with an exclusive two-week free window through Oct 25. No credits deducted, domestic or international version, both. I've been a fan of StepFun for a while. They keep grinding on actual office work while everyone else chases coding benchmarks. The model itself: 600B MoE, only 27B active, 1M context. It scored 44 on the AA intelligence index, good for top 3 among open models worldwide, and it's sitting at #1 on OpenRouter's call volume chart. Deep research, retrieval, data analysis, docs and slide decks, that's its turf.

freebuff, and the models are actually free.

There are more no-cost LLM sites than ever. freebuff's model is simple: text ads pay for inference, you pay nothing. No subscription, no credit card, not even an API key. Billing isn't per token either, it's per one-hour session. Each model has a session price, and once you're in you can send as many messages as you want that hour. Way nicer than watching a token counter tick down. The free tier gives you 25 Freebucks a day, about $0.25, and it resets the next day instead of rolling over. Same budget, wildly different runtime depending on the model you pick: DeepSeek V4.1 Flash is roughly 1 hour a day per the official numbers, and Solar Mini 4 and Solar Pro 4 are listed as unlimited for now. The caveats, because there are some. Outside supported regions, or if they detect VPN/proxy, your daily budget drops to 20 Freebucks and the model list shrinks too. The ToS says your data may be used for training. Under heavy load some models might serve you a quantized version. It's also still beta, so prices and quotas move every few weeks, and their own fine print says sessions can get paused for "fair use" even while a free account still has balance. 🔗 https://freebuff.com/?ref=ref-b22b7bc3-a5cb-4e7c-8446-995e9d6c6dfe

2026-10-14

Muse registration is getting stricter and stricter.

Made another account to help run my overseas social media and content distribution. A lot of cloud browsers block the Muse homepage outright now, so most of the old tricks are dead. The two I can confirm still work: Gemini's Sprak cloud computer, and the built-in cloud browser that ships with Grok premium and up (not Grok-bot's cloud computer. The bot throws you into the waitlist even if you turn auto timezone off).

2026-10-14

Claude appeal rejected.

Won't refund the subscription either. I'm half convinced the appeal email is itself phishing, because I'm pretty sure they log the IP. DHH came to Shenzhen recently and showed how he uses Claude Code over there. That's the safest way now, honestly: run Claude Code on an overseas machine, connect through Tailscale, never use it locally. So I need to rent an overseas box with a residential IP, and register and log into Claude there.

2026-11-14

Sharing my current mobile only workflow.

Until late September it was the usual setup: parked in front of my laptop, terminal open, bossing pi around, fanning out subagents, then eyeballing diffs myself and filing the PR. Glued to the chair all day, every day. Late September I moved the whole thing to my phone, Hermes runs the show, pi takes the orders. My back is very grateful. Three weeks in, still no regrets. Stack: Hermes on flagship brains (gpt-6-astra / opus-5.5), pi on mid-tier (ds-v4.1-flash / sonnet / terra) with as few plugins as possible, the two talking over a2a. Hermes reads the issues, slices them into small tasks, spins up a worktree for each one, and dispatches a pi into it. When a pi finishes it pings Hermes, who decides: ship the PR or bounce it back for rework. I get back from a full day of being a tourist, or just wake up, and a dozen PRs are sitting there done. Pretty sweet. The catch is how black-box the whole thing is. My emotional arc has been a rerun of early vibe coding, except back then I at least read the diffs. Now even that goes to a flagship model. Boxes all the way down. If you don't have data anxiety, grok-bot and chatgpt-dot basically ship this out of the box, grok-cli and codex built in, native agent-to-agent comms. Downside: your code lives on their servers, everything's moderated, and one ban takes it all with it.

2026-11-14