Here's a new model to quant. Meituan just gave us LongCat-2.5.
This is a dev post classified by AGK Intelligence as AI dev tools (a launch), kept by the Dev Radar because it carries real work, not commentary.
Here's a new model to quant. Meituan just gave us LongCat-2.5. Stats 🧠 1.6 TRILLION parameters ⚡ ~48B active/token 📚 1M context 👁️ natively multimodal 🤖 built for long-horizon agents ⚙️ W4A4 INT4 inference Check out the last bullet. Raw weight math calcs 16-bit → ~3.2TB 8-bit → ~1.6TB 4-bit → ~800GB 3-bit → ~600GB 2-bit → ~400GB Yes. Even at 4-bit we're talking roughly 800GB of weights. Meituan says its W4A4 implementation improves 📥 prefill → +43.3% ⚡ decode → +42.8% ⏱️ TTFT → -15.9% versus W8A8. And the jump from LongCat-2.0 is big. Terminal Bench v4: 2.0 → 3.0 2.5 → 38.5 AutomationBen
Posted by David Hendrickson (12.6k followers) 1 h ago · 8 likes · 595 views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- What if you connected a coding agent to your browser and let it catch, fix, and re-test… — @ChromiumDev
- Stop copying the same skills into every coding agent. — @DanKornas
- DeepSeek is promoting Creator Mode, that lets its open-source Harness agent build… — @rohanpaul_ai
- As a dev, I am bullish on this new type of system-one extraction model. — @omarsar0
- "When the voice moment happens, it will be just as, if not more significant than… — @Vapi_AI
- 😱逆天了！12GB显卡就能跑！7630亿参数压到1 bit，DeepSeek V4.1 Flash 1017 最新魔改版来了！专为代码生成 + Agent工作流优化！ — @NFT_Chen
- Maya v1.1.50 is out 🚀 — @PeasantSmith
- TRENDING 📈 @karpathy built HackerTube over a weekend: Claude filtered 43M Hacker News… — @thestartpod
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.3k posts from 5k X accounts over the last 21 days, 9k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-02 18:08 UTC. Full method.