Kir Gz (L/0)

Endlessly exploring reality. Obsessed with modeling futures. Researcher, Engineer, Co-Founder & CTO of Standi. Big dreams: automated companies & science & math

[Chubby♨️ @kimmonismus]

GPT-5.6 Sol on 750 Token/s in Blender.

Not sped up. Holy moly. That’s so much more impressive than looking at benchmarks

[Kir Gz (L/0) @K22R9]

@thekitze It’s several times cheaper on Contabo, server auction. 256gb ram dedicated for $70-90

[Kir Gz (L/0) @K22R9]

I think we should compare effective LLM prices, not some theoretical/nominal. So, when you fully utilize $200 codex subscription, you get ~$14K API equivalent. Also, each reset means +3.5K each. If you have to resets/mo, it means that you have about 21K/mo value, more than 99% discount. So, effective price of gpt 5.6 is ~$0.28/M output tokens. Exactly as Gemini V4 flash

[Kir Gz (L/0) @K22R9]

@NousResearch BYOK? Connect OpeAI subscription?

[Kir Gz (L/0) @K22R9]

@PaulRBerg That is pretty easy to customize and fix:

[Kir Gz (L/0) @K22R9]

@NotionHQ Add BYOK please!

[Kir Gz (L/0) @K22R9]

@ThePushkaraj Please add screenshots, UX animations

[Kir Gz (L/0) @K22R9]

@robj3d3 @levelsio Good options that I personally use:
1. TBC (Georgian, nice cashback, free lounges in airports)
2. Bybit (International, great cashback, fully remote, + QR codes in Vietnam)

[Kir Gz (L/0) @K22R9]

@robj3d3 Btw Da Nang is one of the best places to do that in quality/cost ratio, much better than Bangkok

[Kir Gz (L/0) @K22R9]

@outsource_ Claudafinil, highly addictive, $200 per blister

[Kir Gz (L/0) @K22R9]

@thsottiaux @kr0der speed 500-700+ tps, heavy use of subagents + dynamic workflows like in Claude

[Kir Gz (L/0) @K22R9]

@fchollet Attention

[Kir Gz (L/0) @K22R9]

@jxnlco Brainstorming with gpt 5.5 pro

[Kir Gz (L/0) @K22R9]

@claudeai It's just... worse and more expensive than Opus 4.8

[Kir Gz (L/0) @K22R9]

@itsolelehmann claude traveled 6 months and got restricted by export control

[jietang @jietang]

GLM-5.2 is Fully Open, Frontier Intelligence Belongs to Everyone

Today, the sudden restriction of certain frontier models is deeply regrettable. At a time when access to frontier models is abruptly cut off for non-technical reasons, we are even more convinced of one thing: science should be global.

The path to AGI (Artificial General Intelligence) must never be enclosed by high walls. We have always believed that AGI should be the cornerstone for all of humanity to collaboratively explore the boundaries of intelligence and solve complex challenges, rather than a privilege monopolized by a few rules and subject to revocation at any moment. In the face of external blockades and restrictions, our attitude is one of radical openness. Frontier intelligence must remain open-source, accessible, and buildable, serving every dedicated developer.

GLM-5.2 is Zhipu's most capable open-source model to date. It not only supports a truly usable 1M context window but also maintains a continuous lead in the independent completion of long-horizon tasks, providing solid foundational support for building complex agent applications. It also continues to be our main engine for creating the strongest domestic coding model.

Tonight at 5:21—at this special moment—GLM-5.2 will officially be available to all GLM Coding Plan users (including Lite / Pro / Max). The API will also go live next week.

A step closer to frontier intelligence for everyone.
The future of AI is open, and it is for the people.
ModelKey: GLM-5.2

[Kir Gz (L/0) @K22R9]

@teortaxesTex 52-54

[Kir Gz (L/0) @K22R9]

@pika_labs That's really a new level

[Will Manidis @WillManidis]

Dario (48 hours ago): “US gov should be able to block model deployment”

USG: *export controls models*

Dario: “not like that”

[Kir Gz (L/0) @K22R9]

@OpenAI 5.5 spark?:)