vedant

aligning humans & machines @

[vedant @vedantdotrs]

IT’S ONLY 320B????

[Z.ai @Zai_org]

Introducing GLM-5.3: Built to Code. Ready for Cyber Defense.

- Top-tier coding and agentic capabilities, achieved through post-training on the 743B base model
- A major leap in cybersecurity, setting a new standard among open models

Tech Blog: https://t.co/ekQkO83jCv

[vedant @vedantdotrs]

@gabriel1 IMO it's a balance of "power user" and "good feedback user". Some of our initial users just kept complaining without using the product much, those are the ones you should churn yourself.

[vedant @vedantdotrs]

@SteveTheDust yo steve 😭😭😭😭

[vedant @vedantdotrs]

@ethangwang @0xCL4R Dude you gotta pu to the office again soon

[vedant @vedantdotrs]

@0xCL4R @The_Forty_Four Bro deadass had his shirt auto-tucked 😭 😭 im crying bruh

[vedant @vedantdotrs]

hey @interaction, i think you might be a bit late..

[vedant @vedantdotrs]

I still believe general harnesses can be a thing (e.g. Cursor), but there certainly cannot be one model for all.

The recent taste regressions in "better" models have been the evidence.

[vedant @vedantdotrs]

@erikdunteman imo the @browserbase team is coming up with some bangers too. cloudflare's the pioneer though

[vedant @vedantdotrs]

Qwen and Kimi going band for band.. It’s genuinely over for closed source labs

[Qwen @Alibaba_Qwen]

📢Meet Qwen3.8-Max — our most capable model to date.

Next week, the open weights of Qwen3.8-Max will be released, and Qwen3.8-27B is also going open-weights to meet you all!🎉

Qwen3.8-Max, a new bar for coding and cowork at 2.4T parameters:

- Autonomous coding: 10+ days of self-evolving development, from empty folder to production without hand-holding, complete project trace in the GitHub：https://t.co/iVHZWQoeSo
- Real work, real results: Production-quality deliverables across hundreds of professions.
- Long-horizon mastery: System-level autonomous planning with closed-loop adaptive learning, driving 500+ turns of chip design optimization and 365 days of e-commerce strategy.
- Native multimodal intelligence: Vision isn't just input — it's a continuous feedback loop for planning, execution, and self-correction.

💰Pricing:
Input: $2.0 / M tokens
Output: $6.0 / M tokens
Implicit Caching: $0.25 / M tokens

Start building with Qwen3.8-Max! 🚀

📖 Blog: https://t.co/iwjmQxLBof
✅ Qwen Studio: https://t.co/4V2pFvDovG
⚡ API: https://t.co/gAGqaLQGbN

[vedant @vedantdotrs]

@Alibaba_Qwen $2/$6 in/out for a model better than 5.6 Sol???? Yeah it’s wraps for closed source companies

[vedant @vedantdotrs]

Today is the day I remove all model routers from my harnesses. One model is truly enough: GPT-5.6-Luna:medium.

[OpenAI @OpenAI]

We are committed to pushing the model frontier across cost efficiency, capability, and speed.

Starting today, we are reducing prices for GPT-5.6 Luna by 80% and GPT-5.6 Terra by 20% , and offering a faster option for GPT-5.6 Sol in the API.

Luna and Terra’s lower prices are reflected in how usage is counted in Codex and ChatGPT Work, so your usage goes further.

[vedant @vedantdotrs]

@hazhubble Gonna use this everyday, love it. Congrats, Haz!

[vedant @vedantdotrs]

@waynesutton @convex cant have that without the biggest fan of convex

[vedant @vedantdotrs]

@sksuvorovas Exactly! There's so many companies out there doing domain-specific agents in different industries because general agents that can do domain-specific things are just not possible yet

[vedant @vedantdotrs]

Agentic design is the only thing LLMs are bad at

[vedant @vedantdotrs]

@ZuriObozuwa Lmao I swear it always thinks of things in dev days and not “ai” days

[vedant @vedantdotrs]

@ethangwang 😭😭

[vedant @vedantdotrs]

well this just massively reduced my burn rate (was at SUS for like 5 minutes)

[vedant @vedantdotrs]

@theo dude just use cmux

[vedant @vedantdotrs]

@0xCL4R @therosieum @cpkawesome fire read this literally summed up the entire propaganda machine lol