We just open-sourced THX-01, a 322M decision model that beats Jev on several benchmarks and runs on your phone!
Fully open-source.
Free API: api.hal-x.ai/docs/thx-01
Model weights: huggingface.co/doofz/THX-01
pip install thx01
doofz11 retweeted
We beat TypeSafe Jev on our 2,843-ticket benchmark:
THX-01 98.4% vs Jev 97.4%
Messy text: 98.8 vs 96.2
~10 ms vs 331 ms
And ours is open source. And now the API is free.
huggingface.co/doofz/THX-01
api.hal-x.ai/docs/thx-01
pip install thx01
THX-01 API is now free.
No key. No sign-up.
Just curl it.
200 decisions per minute.
322M open model,
~10 ms per answer, calibrated probabilities.
curl api.hal-x.ai/v1/systemone
docs: api.hal-x.ai/docs/thx-01
Weights: huggingface.co/doofz/THX-01
THX-01 doesn't just classify text. Give it a document and your questions, and it returns structured JSON: numbers, quotes, and the exact passage behind each answer.
Fully open source. Runs on CPU.
huggingface.co/doofz/THX-01
or
pip install thx01
We just open-sourced THX-01, decision model like Jev that outperforms it and runs on phone!
322M params. One forward pass. ~10 ms.
Matches Claude Sonnet 5.5 on our 4-language ticket benchmark, 150x faster.
Apache 2.0.
huggingface.co/doofz/THX-01
doofz11 retweeted
🧩 DeepSeek Harness v0.1 is now available in Developer Preview!
🔹 We’re opening it up to developers building agent harnesses worldwide and open-sourcing the codebase in MIT license.
🔹 Powered by the Cordis meta-framework, DeepSeek Harness is an agent harness built around one core idea: Everything is a plugin. Models, tools, skills, sessions, sandboxes, filesystems, loops, orchestration, and UI are ALL implemented as plugins, and can be mixed, matched, replaced, and extended.
Try it now!
github.com/deepseek-ai/deeps…
doofz11 retweeted
BREAKING: Dario declares open models may be conscious and too dangerous to download. urges the Pentagon to intervene. citing grave national security concerns that americans can currently download frontier ai for free.
doofz11 retweeted
We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model.
To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection. Moonshot AI has also acquired GB300-equipped servers and has accessed GB300s in Thailand, likely to train its AI models.
The United States strongly supports the free and fair development of AI, including a thriving competitive ecosystem that spans frontier models, specialized systems, open-source frameworks, and open-weight models. Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.
doofz11 retweeted
dont do this btw
744B parameters. On a laptop. With 25GB RAM.
Colibri runs GLM-5.2 (744B MoE) in pure C with zero dependencies. The trick: only ~40B params activate per token, so it keeps the dense part resident and streams experts from disk on demand. A single 2,400-line C file. No GPU, no BLAS, no Python at runtime.
This shouldn't work. But it does.
⭐ 2.1K #AI #OpenSource
github.com/JustVugg/colibri
Follow for daily dev finds 🔔
