Open source

AIThe skills.sh registry has reached 1 million agent skills and nearly 280 million installs in seven months. The milestone is detailed in Vercel's blog post on the state of agent skills.

AITogether AI says teams can access frontier open models through its API without managing underlying infrastructure, a point Ted Cui, its VP of Engineering and Inference Platform, made at Apsara Conference 2026. The company emphasizes reliable, fast inference as essential for this access.

AIvLLM has published a blog post explaining how its integration with TileRT works, alongside a public benchmark configuration in the InferenceX repository. The benchmark script covers GLM-5.3 FP8 on MI355X hardware and is linked on GitHub. The post itself provides the integration details.

AIThe TileRT and AMD teams reached 469 tok/s single-user decode for GLM-5.3 on 8× MI355X using vLLM. The setup disaggregates work, with vLLM handling prefill and TileRT handling latency-critical decode through vLLM's V1 connector interface. SemiAnalysis's AgentX benchmark reports the configuration at 470 TPS on GLM 5.3 (FP8), over 40% faster than GB300 TRTLLM using FP4.

AIHugging Face released SmolDataEnvs, more than 5,000 open-source RL environments aimed at real-world data science tasks. They target the gap between simple educational games and frontier-level benchmarks, especially for improving coding in models under 10B parameters. The environments are designed as a testbed for developing new RL methods such as GRPO or OPSD.

AILangChain launched LangSmith Fine-Tuning, which lets users fine-tune open models on their LangSmith agent traces using the open-source smithtune CLI. Training runs on Baseten Loops in the user's own workspace, and smithtune deploy places the evaluated checkpoint on a Baseten Dedicated Inference deployment. Loops is in early access, so users may need to request access for their workspace.

AIGitHub Security Lab describes the Fuzzing Taskflow, an LLM agent pipeline that identifies entrypoints, writes harnesses, runs AFL++, reads coverage reports, and triages crashes for C/C++ repositories. The agent makes decisions while MCP tools handle execution, and state is stored in a SQLite database. The post also warns that the taskflow runs AFL and build commands directly on the host, so it should be used only in disposable environments without elevated privileges.

Why it matters: The post explains how an LLM agent automates fuzzing steps like harness writing, coverage gap chasing, and crash triage, with a runnable workflow and design tradeoffs.

AIOdyssey launched Agora-2, a multi-agent world model that the company is making available for public experimentation. The post predicts such models will increasingly power applications in AI training, AI safety, robotics, autonomous vehicles, defense, energy, cybersecurity, and gaming.

AIThe RLKernel team integrated RL-Align/RL-Kernel with vllm-project/vime, and a 200-step Qwen3-8B GRPO run on 8× AMD MI300X recorded zero logprob mismatches between Megatron training and vLLM rollout. The strict path aligns reduction order, intermediate precision, rounding points, and math primitives across both sides to achieve bit-for-bit matching on ROCm.

AINVIDIA shared a blog post linking to a story about an open protein dataset, tagging Google DeepMind and EMBL-EBI. The post itself provides no further details about the dataset's contents, scale, or availability.

AIGoodfire Research introduces Block-Sparse Featurizers (BSF), which decompose model activations into subspaces rather than single directions. Applied to DINOv3 and Stable Diffusion XL, BSFs find interpretable multidimensional features that better explain activations and enable fine-grained steering. The authors report that most concepts they examined have a stable rank of about two to four dimensions.

AIMind Lab unveiled Mint Recursive, a post-training and inference platform for industry use, alongside Macaron-V1.1, a model post-trained entirely on it. Macaron-V1.1 is a 752-billion-parameter model built from GLM-5.3 with four two-billion-parameter LoRA expert modules for chat, agents, coding, and generation. The platform is serverless and bills by token usage, and it collects feedback from models in use to support continued training.

AIinclusionAI has published public training-content summaries on Hugging Face for its Ling and Ring model versions, including Ling-2.0, Ling-2.5, Ling-2.6-1T, Ling-3.0, Ring-2.0, Ring-2.5-1T, and Ring-2.6-1T. The documents, organized under the template associated with Article 53(1)(d) of Regulation (EU) 2024/1689, contain documentation only, not model weights or training datasets. Each summary covers only the model versions it names.

AISupporting plugins can call an optional model for structured choices, kept separate from chat. TypeSafe Jev sends supplied information to its hosted API and incurs normal charges, while ONNX offers local CPU options. Decision Models are off by default.

AIOpenClaw lets users keep public GitHub issues, PR discussions, commits, and diffs open beside their chat. Several references can be followed in tabs. The feature is read-only and public-only, so private content and edits remain on GitHub.

AIMidjourney's alpha site now lets users preview their current prompt across styles with "Live previews" in the Styles sidebar and save prompt-bar settings as defaults via Settings → Advanced → Your defaults. The Create feed received a full-width masonry redesign with hover-based prompts and buttons, and Korean is now live for all users on midjourney.com.

AIAlex Wortega releases openjev, a family of Jev-style models at 0.8B, 2B, 4B, and 35B-A3B for reranking, grading, guardrails, and real-time game decisions. The models come with SGLang serving support, according to the post.

AILiquid AI released an experimental DSpark draft model for its LFM2.5-VL-3B vision-language model, delivering decoding throughput gains of up to 2.66× on GPUs and 3.13× on edge devices. The drafter adds about 280M parameters, an 8.9% increase in the deployed model's parameter count, and is available on Hugging Face with support in llama.cpp, SGLang, and MLX-VLM.

AIvLLM now supports Gumbel-max watermarking, which embeds a keyed signal into generated text without changing the expected token distribution. Detection requires the secret key and tokenizer, and the signal accumulates over longer outputs. Benchmarks on Qwen3.5-27B with MTP-3 show throughput changes between -1.1% and +2.0% across batch sizes, with no consistent slowdown.

AIGoogle Developers reproduced Ai2's Olmo 3 7B from scratch in MaxText on Google Cloud TPUs, covering both the stage-1 pre-training run and the stage-2 mid-training anneal. The match was checked on held-out C4 loss, an 8-task accuracy suite, multi-domain perplexity, and token-level KL, not just the training loss curve. The post also describes a data-loader bug that made training loss look better than the reference while held-out metrics did not move.

Why it matters: The post documents how a faithful reproduction was verified on held-out metrics, including a data bug that training loss alone would have hidden.

AIInferact says vLLM maintainers used megakernel optimization to reach 700 tokens per second per user on TPUv7 running Kimi K3. SemiAnalysis, which shared the work, reports this is 56% better performance than Nvidia's GB200 NVL72. Inferact links a full technical breakdown of the TPU megakernel work on its blog.

AIBlack Forest Labs has published a blog post and downloadable weights for its FLUX 3 Action model, with links to a Hugging Face collection. The post also invites readers to contact the company's robotics team.

AIJulien Chaumond of Hugging Face posted a thank-you message to supporters of arXiv. The quoted arXiv post reports a $17.2 million multiyear investment from Simons Foundation International, Siegel Family Endowment, and XTX Markets.

AIGoogle's Gemini app now connects with 13 new apps, including Adobe, Squarespace, and Peloton. Josh Woodward asks users which favorite apps they want added next.

AIMicrosoft Research reports that running physical AI inference on onboard GPUs can limit robot performance and battery life, while offloading inference to edge or cloud GPUs improved results in mobile manipulation tests. In its evaluation, smaller onboard GPUs slowed mapping and planning by up to 383% compared with an A100, and large onboard GPUs such as Jetson Thor drained robot batteries by up to 160%.

Why it matters: The study measures how offloading robot inference to edge or cloud GPUs changes task success, battery life, and model size, offering evidence for infrastructure design.

AIUnsloth says local AI adoption in 2026 has grown far beyond its own predictions. The company also reports surpassing 500M model downloads on Hugging Face, with Qwen3.8-27B GGUF as its most-downloaded model ever. It says it is working on new features to make local AI faster and more accessible in the coming weeks.

AIMike Knoop says the top ARC Prize 2026 ARC-AGI-2 score of 83.06% by Tufa Labs is only 2% short of the 85% grand prize threshold. The challenge runs under strict Kaggle compute limits with no internet access, and the winning solution is set to be open sourced. The image shows the leaderboard with RabbitHole at 76.94%, nvbanana at 74.17%, Yi-Chia Chen at 55.14%, and Kha Vo at 37.50%.

AILewis Tunstall of Hugging Face shares his first cameo as a voice actor using the Reachy Mini robot, linked to a YouTube video about Australian football. The post itself provides no technical details, and the Reachy Mini context comes from a separate post by Andi Marafioti about NVIDIA's open-source Nemotron 3 Diarization model.