THX-02 from @halx_ai is open source: a 4.5B multimodal decision model trained on…
This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
THX-02 from @halx_ai is open source: a 4.5B multimodal decision model trained on requests up to 32K tokens, with a 60M adapter for selective thinking. It returns probabilities over your supplied options. 🤖 https://modelscope.ai/models/HAL-X/THX-02 🧠 Reason selectively: Auto mode generates a short reasoning note when the direct answer is uncertain and an internal gate identifies a need for reasoning, then combines the direct and thought-assisted predictions. ⚡ Fast direct decisions: Median latency is 50 ms directly and 79 ms in auto mode on one RTX PRO 6000. Requests that trigger thinking typ
Posted by ModelScope (16.8k followers) 2 h ago · 5 likes · 467 views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: modelscope.ai.
More dev work like this
- My fav bit about our Neon MCP usage numbers: You can see model launches in the client… — @andrelandgraf
- Introducing Mobile Diffs. — @linear
- GitHub Copilot can now generate images! — @pierceboggan
- Already trained Ultralytics YOLO models in W&B? Move them to the Ultralytics Platform… — @ultralytics
- Meet Gemini Omni 1.2 Flash ⚡️ — @GoogleAI
- Gemini Omni 1.2 Flash is now live on Segmind. — @Segmind_ai
- You can't make your agents faster if you can't see what's slowing them down. — @mastra
- Introducing P-Video-2-Pro-Replace, based on @MiniMax_AI H3! 🎬 — @PrunaAI
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. 31.6k posts from 5k X accounts over the last 21 days, 9.4k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-05 18:08 UTC. Full method.