𝕏   Ahmed  𝕏

[𝕏   Ahmed  𝕏 @eng_ahmd]

@Alibaba_Qwen @arena Still waiting for the MoE ⏳

[𝕏   Ahmed  𝕏 @eng_ahmd]

Oh ya FreeToken is beast 🎉🎊

[𝕏   Ahmed  𝕏 @eng_ahmd]

This bad boy called FreeToken is sick I can’t believe running NVFP4 on windows with my rtx 5080 GPU speed I was never able to get earlier give it a try 💥💥💥🤖💻

[𝕏   Ahmed  𝕏 @eng_ahmd]

@Andy_ShuoYang This is sick! I love it on the first glance. 💥

[𝕏   Ahmed  𝕏 @eng_ahmd]

This bad boy called FreeToken is sick I can’t believe running NVFP4 on windows with my rtx 5080 GPU speed I was never able to get earlier give it a try 💥💥💥🤖💻

[Shuo Yang @Andy_ShuoYang]

Your gaming PC can now serve frontier models at interactive speed using official checkpoints without extreme quantization!

Qwen3.6 35B → 8GB RTX 4060 laptop @ 39 tok/s

DeepSeek-V4-Flash 284B → RTX 5090 desktop @ 22-25 tok/s

GLM-5.2 753B → RTX PRO 6000 workstation @ 15 tok/s

Run your claude code or codex now with frontier model for $0

Meet FreeToken 🧵

[𝕏   Ahmed  𝕏 @eng_ahmd]

--reasoning-budget-message "... Time to stop thinking. Give the final answer or make the tool call now." --spec-type draft-mtp --spec-draft-n-max 3 --spec-draft-p-min 0.80 --top-k 20 --top-p 0.95 --min-p 0.0 --repeat-penalty 1.0 --kv-unified

[𝕏   Ahmed  𝕏 @eng_ahmd]

Thanks to @UnslothAI I am able to run @Alibaba_Qwen 3.8 27B dense model on rtx 5080 laptop 16gb VRAM with up to 50 tok/sec with context window 90K this is insane running it with hermes agent it works like a charm so far I am so happy with it but will keep pushing the limits

[Unsloth AI @UnslothAI]

We’re releasing new Qwen3.8-27B GGUFs with 10% higher accuracy.

Unsloth Dynamic V3 outperforms others by >10% on Div-300, KLD & more benchmarks.

We also release 1-bit quants that retain 77% accuracy. Run on 8GB RAM.

Blog: https://t.co/tHsBexyh2K
GGUF: https://t.co/xIdNwm7CLQ

[𝕏   Ahmed  𝕏 @eng_ahmd]

--cache-type-k q8_0 --cache-type-v q8_0 --reasoning on --image-min-tokens 1024 --cors-origins localhost --chat-template-kwargs "{"preserve_thinking": false, "reasoning_effort": "low"}" --mmproj mmproj-F16.gguf --reasoning-preserve --reasoning-budget 4096

[𝕏   Ahmed  𝕏 @eng_ahmd]

@UnslothAI @Alibaba_Qwen GOAT 🐐

[𝕏   Ahmed  𝕏 @eng_ahmd]

@Alibaba_Qwen @cline when the 35B MoE ?

[𝕏   Ahmed  𝕏 @eng_ahmd]

@MiaAI_lab @Alibaba_Qwen +1

[𝕏   Ahmed  𝕏 @eng_ahmd]

@NYCMayor @TDFNYC Mayor, last week an NYPD officer ticketed me for a clear license plate cover (legal in my state btw) with no warning. That same day I witnessed a driver run a red light with no officer present. Which issue should officers prioritize to keep NY safe? I am not coming to NYC again!

[𝕏   Ahmed  𝕏 @eng_ahmd]

@UnslothAI @Alibaba_Qwen You guys are the Local AI community heros 🥰🙏💥

[𝕏   Ahmed  𝕏 @eng_ahmd]

@Alibaba_Qwen Insane 💥🥰

[𝕏   Ahmed  𝕏 @eng_ahmd]

Proud to be one of the first few hundred of the 500 contributors. This AI has 10x’d my productivity it’s not just a tool, it’s a co-worker, executive assistant, and timekeeper in one. Huge thanks to the Amazon community. Any business function can use it not only SDE 👻♥️

[Darko Mesaroš @darkosubotica]

The amazing folks from @kirodotdev have delivered something delightful! An Open Source development tool that wrangles a bunch of complexities of being a software developer! Welcome Kiro Crew! ❤️ Check out the Git repo down below! 👇

https://t.co/UEOgN4P09e

#MyKiroCrew

[𝕏   Ahmed  𝕏 @eng_ahmd]

@Pokee_AI How to deploy it locally if it is closed source !

[𝕏   Ahmed  𝕏 @eng_ahmd]

@QwenDevs @Web3Twon Indeed MoE is badly required 🤖

[𝕏   Ahmed  𝕏 @eng_ahmd]

@Alibaba_Qwen What about MoE !

[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

[David Ornside @David_Ornside]

🚨🗣 Jamie Carragher on Egypt's Canceled goal by VAR:

🗣"I am telling you that, if that was against another team it would have been awarded as a goal"

"If that was in the premier league, Laliga or Serie A, it would have been a goal even after VAR review, there is a lot of inconsistency lately in this tournament"