Join free with your email
Free. No spam, ever — we'll never share your email address and you can opt out at any time. Already a subscriber? Log in
What credible people across AI noticed, why it matters, and where the field is converging or disagreeing.
Posts are grouped by conversation and tone. Select a lens to filter the stream; these are never permanent labels on people.
One card per development. Sources are clustered; reaction bundles describe these posts, never the people behind them.
2 directory members surfaced this signal.
“Hydrogen 1.0 🔬 Impressive open model built for scientific discovery huggingface.co/deep-princip... - Based on GLM 5.3 - 1M context window - MIT license + with BF16, NVFP4,MXFP4 variants”
2 directory members surfaced this signal.
“Kimi K3.1 Pro 🔥 huggingface.co/moonshotai/K...”
2 directory members surfaced this signal.
How teams are shipping and applying it.
“SparkLLM's Spark-X2.6-4B (open-weight) - 4B + Apache 2.0 - Up to 1M context - 3 thinking modes - Local deployment via GGUF Model: huggingface.co/collections/...”
Background, chronology and why it matters.
“After its 4B model, SparkLLM is back again with an even smaller one: Spark-X2.6-1.7B huggingface.co/collections/... - 3 thinking modes: none, low & high - Up to 1M context - Apache 2.0 + GGUF/int4/int8/FP8 - Leading decoding throughput vs similarly sized mo…”
2 directory members surfaced this signal.
“GitHub: github.com/baidu/Unlimi... Hugging Face: huggingface.co/baidu/Unlimi...”
“Unlimited OCR 2 🔥 New 3B OCR from Baidu huggingface.co/baidu/Unlimi... - MIT license - Unlimited horizon document parsing - SR SWA attention with nearly constant per step memory & latency - SOTA on 100 page MarathonDocBench”
2 directory members surfaced this signal.
“Sage A1.5 🔬 New agentic model built for long horizon scientific problem sloving huggingface.co/collections/... - 35B MoE - 256k context - Parallel subagents - Built in self-verification”
1 directory member surfaced this signal.
“ERNIE-5.5-Flash-Base 🔥 New open MoE model from Baidu huggingface.co/collections/... - 280B / A 14B - 33.5T pre-training tokens - Up to 1M context - Hybrid sparse attention - Raw base checkpoint, no post training - Custom license”
1 directory member surfaced this signal.
“SenseNova-Loop-9B-Preview 👀 A looped multimodal model from SenseNova huggingface.co/sensenova/Se... - 9B - Apache 2.0 - 56 layer executions via weight sharing - Up to 256K context - Hybrid attention + two stream mHC”
1 directory member surfaced this signal.
“✨ MiniCPM OCR Jev - 2B - Apache 2.0 - Evaluates OCR accuracy, even from other models - 0.936 AUROC on JevOCRBench - Helps flag errors for human review huggingface.co/collections/...”
1 directory member surfaced this signal.
“Hy SimulMT 🎧 New real time simultaneous translation model from Tencent huggingface.co/collections/... - 30B/A3B - Apache 2.0 - Support Chinese <>English - 3s lag, lowest in its RealSI eval - GGUF & fp8”
1 directory member surfaced this signal.
“Ant group keeps shipping 🚢 They just released weights for Ling-3.0-flash-Sante 🩺 A new open medical reasoning model huggingface.co/inclusionAI/... - 124B/5.1B active MoE - 256K context - MIT License - Agentic deep research & tool use”
1 directory member surfaced this signal.
“StreamEval 🎬 A new benchmark for real time interactive video AI huggingface.co/datasets/Str... - 300 test cases (15–300s) - 5 tasks: avatars, world generation & live editing - 37 models evaluated”