Nvidia PAIR is a free piece of software that finds the computers already sitting on your home network and puts the idle ones to work on AI inference. Announced on 3 September 2026 at IFA in Berlin, the Personal AI Router turns a gaming desktop, a work laptop and a recent Mac into what Nvidia […]
Tencent open-sourced Hy4 preview on 28 August 2026 under Apache 2.0: a 770B Mixture-of-Experts model that activates just 49B parameters per token and reads a one-million-token context window. This breakdown covers the full architecture, from 256 routed experts per layer to Gated DeepSeek Sparse Attention and the built-in speculative decoding layer; every benchmark figure Tencent published, including the 163-expert blind evaluation against GLM 5.3 and Kimi K3; the API price list against GPT-5.6 Sol; the eight-GPU serving recipes; and the four caveats worth naming before any of it reaches production.
Veracity is an open-source fact-checking system from Montreal’s Complex Data Lab that pairs a large language model with a web retrieval agent and refuses to answer without showing its evidence. Every claim comes back as a reliability score between 0% and 100%, a plain-English interpretation, a share recommendation gated at 60%, and a panel listing every source consulted along with its documented credibility. This breakdown covers the five-stage pipeline, the stack behind it, why the 60% cutoff is the most questionable decision in the design, what independent benchmarks say about the roughly 83% real-world accuracy ceiling for automated fact-checking, and how to evaluate any claim-scoring layer before you let it near a workflow.
Alibaba open-weighted Qwen3.8-Flash-Next on 26 August 2026: a multimodal mixture-of-experts model with 125 billion parameters, a separate 51-billion-parameter N-gram embedding table, and just 6 billion parameters activated per token. This breakdown covers the four rebuilt subsystems — Gated DeltaNet paired with Qwen Sparse Attention at block granularity, a Gated Residual stream widened to four gated branches, the N-gram table that offloads to host RAM, and the Muon plus AdamW training recipe with batch-size warmup removed — alongside the 48-layer stack of 512 experts that fires eleven per token, the published benchmark table showing 62.5 on SWE-bench Pro against 53.4 for Claude Opus 4.6 and 84.5 on AndroidWorld against 62.0, the single loss on Humanity’s Last Exam at 35.9 against 40.0, the unverifiable one-ninth training cost claim, the 262,144-token native context extended to a million with YaRN, hosted pricing of $0.16 and $0.47 per million tokens against $2.00 and $6.00 for Qwen3.8-Max, the real hardware bill from a 172.78 GiB FP8 checkpoint down to a 111 GB four-bit GGUF, the qwen-community-1.0 licence that is not Apache 2.0, and a buyer’s checklist for treating a preview checkpoint as a production dependency.
On 26 August 2026 the mystery ended: Z.ai, the Beijing lab formerly known as Zhipu AI, confirmed that the anonymous Ox Alpha model topping OpenRouter and OpenCode was the newest iteration of its GLM series, and published the weights the same evening as GLM-5.3-Flash under an MIT licence. This breakdown covers what was confirmed and when, the architecture the model card revealed — 320 billion total parameters with just 18 billion active, hybrid sparse and linear attention, a 1,048,576-token context and forced reasoning that cannot be disabled — the published benchmark table showing 84.3 on Terminal Bench 2.1 against 85.0 for Claude Opus 4.8 and 87.4 for GPT-5.6 Terra, the viral 80 per cent DeepSWE claim that came from a 10-task subset and collapsed to 63.4 on the full 113-task run, the 44 trillion tokens and 503,000 users the stealth week generated, the tokenizer and error-code forensics that unmasked the lab before it spoke, the $0.15 and $0.50 per million token pricing, the company’s Hong Kong listing and US entity-list status, and a buyer’s checklist for deciding between the hosted API and self-hosted weights.
Hugging Face — the open AI hub often called the GitHub of AI — is reportedly exploring a sale that could value it at $13 billion or more. Business Insider broke the story on 23 August 2026 and Reuters confirmed a bank is sounding out bidders, though no buyer has been named and the talks are early. This article separates the confirmed claims from the speculation: the valuation arithmetic against the $4.5 billion Series D, the July security breach hanging over due diligence, the plausible buyer scenarios from hyperscalers to chip vendors, what a change of ownership could mean for open-source AI neutrality, and the practical steps businesses that depend on the hub should take now.
The global race to build increasingly capable artificial intelligence models has entered another significant chapter with the release of Kimi K3 AI, the latest open-source large language model developed by Chinese AI startup Moonshot AI. Announced as the company’s most powerful model to date, Kimi K3 AI has attracted worldwide attention for its massive scale, […]
Open-source artificial intelligence continues to reshape the software industry, giving developers unprecedented access to powerful AI technologies that were previously available only through proprietary platforms. One of the latest developments attracting attention is Grok Build, which has been open-sourced on GitHub while also introducing a reset to its usage limits. This move reflects the growing […]
DeepSeek has unveiled a V4 preview with Pro and Flash variants, a 1M context window, and a broader web, app, and API rollout.
A practical guide to AirLLM, the open-source inference library that reduces memory requirements for very large open models by loading layers one at a time.
