TiQTiQ💥

@kimmonismus if you wanna use deepseek models. Url to signup : agentrouter.org/register?aff=r… Make suru you use an OLD github account to signup to get the 150$ bonus balance. make sure to use this reference link to get the maximum balance for you. Creating with this link you get you extra 25$

DeepSeek just made its 75% price cut on V4-Pro permanent. Xiaomi's MiMo slashed V2.5 pricing by up to 99%, effective today. Most coverage frames this as a price war. The more interesting part is the engineering that makes these numbers sustainable. DeepSeek's V4 paper describes a *hybrid attention architecture* that attacks the core bottleneck of long-context inference: the KV cache. Traditional transformers store key-value pairs for every token in the context. At 1 million tokens, this cache alone can fill an entire GPU's memory. V4 introduces two interleaved attention types. Compressed Sparse Attention (CSA) compresses every 4 tokens into a single KV entry, then selects only the top-k most relevant compressed blocks per query. Heavily Compressed Attention (HCA) goes further, compressing 128 tokens into one entry and running dense attention over the result. The compressed sequence is short enough that dense attention stays cheap. V4-Pro's KV cache at 1M tokens is 10% (!!) of V3.2's. Single-token inference FLOPs drop to 27% (!!). The model has 1.6 trillion total parameters but only activates 49 billion per token through Mixture-of-Experts routing, the knowledge capacity of a massive model at the compute cost of one thirty times smaller. MiMo's approach is different but lands in the same place. Xiaomi's team implemented Sliding Window Attention via SGLang HiCache, reducing KV cache data transfer across GPU memory, CPU memory, and SSD to roughly 1/7 (!!) of previous volume. Cacheable tokens expanded by 5x (!!). Combined with expert parallelism optimization and input length bucketing, per-token serving cost dropped enough to make permanent pricing at these levels viable. V4-Pro now sits at $0.87 per million output tokens. MiMo V2.5-Pro at roughly $3/M output, with Flash variants far below that. A year ago, sub-dollar output pricing meant you were using a small distilled model with real capability tradeoffs. These are frontier-class reasoners with million-token context windows. Both companies can commit to permanent cuts because the reductions come from the architecture itself. When your attention mechanism physically processes fewer FLOPs per token and your cache occupies a fraction of the memory, the cost to serve is structurally lower. The price follows the cost curve.

@direkturcrypto hare is a better option . Url to signup : agentrouter.org/register?aff=r… Make suru you use an OLD github account to signup to get the 150$ bonus balance. make sure to use this reference link to get the maximum balance for you. Creating with this link you get you extra 25$

xiaomi mimo offer free $2 after signup. > instant, no credit cards > just signup with email and input referral code after signup Code: V9CAPK ( Enter the code in the console sidebar after signup ) link: platform.xiaomimimo.com/?ref=V9CAPK

@githubstatus I though it was my issue for past 30 minutes 🤦‍♂️

We are investigating reports of degraded performance for Actions and Pages githubstatus.com/incidents/gnft…

@atlasfunded I am hidden

Any comment with the Zero likes on this post after 24 hours will win a FREE $25K instant account 🎉 Winners will be announced tomorrow! ⏰ Good luck!

@0xSero DeepSeek v4 pro OR GLM 5.1 for coding ? Not about local nor price, just want to which one is better for coding.

Deepseek-v4-Flash beats Sonnet, and Opus-4.5 (no thinking) and basically matches GPT-5.2 medium Tomorrow I will have a compression of Flash that'll make it fit well on 1x Spark at hopefully better quality than alternatives Join tomorrow: luma.com/reap

@gitlawb Well deserved 👏🫡

Winners: Top stars before deadline: LawbClicker - $700 Gitsolarstrike - $500 Gitmaps - $500 betcoffeeshop - $400 SafeSpot - $300

Free AL (150$) balance that gives claude Opus-4.6, DeepSeek v4-Pro and Flash, GLM 5.1 Url to signup : agentrouter.org/register?aff=r… Make suru you use an OLD github account to signup to get the 150$ bonus balance. make sure to use this reference link to get the maximum balance for you.

Free AL (150$) balance that gives claude Opus-4.6, DeepSeek v4-Pro and Flash, GLM 5.1. Url to signup : agentrouter.org/register?aff=r… Make suru you use an OLD github account to signup to get the 150$ bonus balance. make sure to use this reference link to get the maximum balance for you.

Url to signup : agentrouter.org/register?aff=r… Make suru you use an OLD github account to signup to get the 150$ bonus balance. make sure to use this reference link to get the maximum balance for you.

@gitlawb @Xiaomi @XiaomiMiMo just want to say thank you @XiaomiMiMo and @gitlawb ♥️

Imagine this running FREE for 2 weeks. Massive shoutout to @Xiaomi and @XiaomiMiMo for backing OpenClaude users. Frontier AI access. Free. Unlimited. Build all you can. AI ACCESS FOR ALL.

@cerebras @cognition Do i have any chance as a student.

🎁 We're giving away 5 Windsurf plans ($250 credit each)! Try SWE-1.6 — @Cognition’s latest fast and intelligent agentic coding model, powered by Cerebras. In a side-by-side with Claude, the speed difference is clear. More iterations, faster fixes, better code. 💬Comment why you want access to enter. Five winners will be selected at random within 48 hours.

@browser_use email sign up not working

One click. Get a 24/7 personal assistant in one click.

@uwukko Can we have 🇨‌🇴‌🇳‌🇹‌🇦‌🇮‌🇳‌🇪‌🇷‌🇸‌ that brave browser is implementing in helium ? Just yuik it 🤭

i opened microsoft edge and it pinned itself to the start of the dock without a prompt or consent running into big tech malware like this makes me appreciate what we do with helium even more

@uwukko Auto update to windows when ?

what if there were native materials in helium... haha jk unless?

@markoilico Give me

If you're now designing or redesigning a website, this will help you a lot. I recently curated the best hero sections, footers, social proof and other website parts because I got tired of having 15+ tabs open (even with Mobbin). Giving it away 100% free. Comment on this post, and I'll send a Figma link to your inbox!

@adeebrq @figma Where is the link/file?

400 followers already??? Fine. I’m dropping the @figma file for my most VIRAL hero section designs, absolutely FREE. - Like + retweet & Drop a comment - Follow me (so I can DM it to you) Only sending it to people who do all 4.

@Trae_ai Fix your sand box for running tests, specially for flutter

TRAE Weekly Highlights ✅ - Important Update: New Membership Plan - New Custom Models: MiniMax-M2.5, GLM-5 - New Feature: Question in TRAE - TRAE Friends and Hackathons: Jakarta, Brazil, Tokyo and Houston TRAE Build: showcasing your projects in Discord and Reddit!

@stitchbygoogle Sheesh! Google is cooking hard this christmas🫡

Shipmas Day 3. Part 2. 🚢🎁 We named our product Stitch for a reason. 🧵 And today, we finally let you do it. Introducing Prototypes. You can now select multiple screens and "stitch" them together into a fully functional, clickable user flow. ✅ Create clickable user flows ✅ Test interactions and animations ✅ Ask for edits by clicking divs on the screen ✅ Export the full context to AI Studio or other coding agents This is Day 1 of interaction design in Stitch, and we have so much more coming. And yes, you can export your entire prototype flow directly to coding agents like AI Studio. We are moving from "Generating UI" to "Designing User Experiences”.

@iamdothash Can we get this for KDE plasman?

Updating Jot for Omarchy with a fuzzy search to make it easy to find your ramblings. ctrl + p yay -S jot-git