Yi

OpenRouter, Carpooling edition. wokey is a marketplace for LLM quota. It pairs people with unused plan credits to people who need more, and takes a 20% cut. Sellers keep 80% of what their credits earn; buyers save up to 90% of the official API price. I use it mostly for flagship models. The models check out as the real thing, cost less than going direct, and I skip the subscription jugglingling. If all you call are cheap models, a router still beats it. 🔗 https://wokey.ai/r/GKT1C78R

Meituan shipped LongCat-2.5. 1.6T params, ~48B active, native multimodal, aimed at long-horizon agents, trained start to finish on Chinese hardware. Images get quantized into multi-level discrete tokens, and that's where Visual Coding and 3D pick up. On the RL side, failed trajectories get turned back into verifiable training tasks. W4A4 inference, 43% faster prefill. Meituan's own evals put it between Claude Opus 4.8 and Opus 5.

SenseTime open-sourced SenseNova Loop-9B-Preview. It stacks layers instead of parameters. The decoder has 32 layers, but layers 4 through 27 run twice with shared weights, so a forward pass does 56 layers of work while the checkpoint stays at 9B. Their own eval numbers: 71.38 on BIRD-Dev and 13.72 on AutomationBench, both above Qwen3.5-27B. Terminal Bench is still behind.

I stuffed an agent into an Android phone. You've probably run into those short videos lately where an agent drives a phone around by itself. This one is called Mika, and it's a free open-source project. It's an Android app that works in the background through accessibility mode, so it can use pretty much any other app you have installed. Reply to someone on WeChat, scroll Xiaohongshu, drop something in a cart and check out. It finds the buttons by looking at the screen. There's no API to wire up, and it doesn't care what resolution your phone is. When it reaches a payment step or a captcha it stops and lets you finish. 🔗 github.com/wangzan101/Mika

My product went profitable, and something in me broke. Someone on Xiaohongshu posted a while back that she sells phone keyboard skins she made with Codex. Eight, ten yuan a pack. Over a thousand a month. I've been at mine for months. October: a few hundred copies, $1,200 gross. Platform fee 30%, tax 10%, so 60% should have landed. It came to just over a thousand dollars. I still don't know how the math pulled that off. I have an overseas bank account. Once lost $500 to a wire transfer off one. I'd still tell you to take payout through WorldFirst: less hassle, lower fees. You can also get a Hong Kong USD credit card tied to it that spends your WorldFirst balance directly. Digital card is free. Physical is $4. Mainland purchases under ¥200 skip the processing fee, and right now there's 1% cashback.

DeepSeek open-sourced dsh-superagent, the agent it uses internally. It stays online, picks up work on its own, files a PR when it's done, and reports back, the way a coworker would. DeepSeek's own design goal is to take as little human attention as possible and hand back as much value as possible. The real shift is going from an agent that sends you messages to one that does the work. Three designs worth copying. Deciding whether to accept a task is split into two layers: a gatekeeper in plain code (is this person on the allowlist, did they @ me), and an advisor, which is the model reading the surrounding context. Cheap checks stay cheap, and the judgment call is the only thing that costs a model call. After the agent takes a job it posts a work badge, stating which version of the rules it is acting on, how far it is allowed to make changes, and who approved it. All of that before it touches code. Concurrency gets a hard cap, but one slot is always held open for messages from people. This is the part I like most. Even with the queue full, you can still redirect a running task or ask for a status update. Without that reserved slot, work piles up and you find out an hour later that it did the wrong thing at full speed.

The official Hermes Agent mobile app has entered public beta on iOS and Android.

Both opencode and commandcode are giving out $100 of ds-v4.1-flash credits. Their $10 plans are going toe to toe. opencode barely finished announcing before commandcode matched it. Don't overthink it. These two are the best value you'll find right now, and if you just go by token cost, it's effectively a 0.1x rate.

My daily token spend is down to around ¥30 now. I used to run the main model only on the official API. Lately I've moved to a relay and honestly it's kind of great: top up at ¥1 for $1 of credit, and ds-v4.1-flash is just 0.15x the price. On top of that, workbuddy, qoder, opencode, codex and the like all throw in free quotas plus some cheap low-tier models. I route those to subagents and the load gets spread across roughly 400M tokens.

Flovia's free API pool is surprisingly deep. The free tier gives you a 5-hour window and a weekly call cap for now, with a paid tier coming later. The ones I'd actually use: Exa, Firecrawl, X, and OpenAI's gpt-6-luna.