Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A developer tracked OpenRouter's daily usage rankings for a month to see which AI models developers actually use in production. The top five models by token volume were all from Chinese labs or open-weight providers, led…
AI-assisted coding tools have sparked a platform shift, with Cursor reaching a $29.3 billion valuation and Lovable hitting $400 million ARR, as 63% of users are non-developers. The cost to build a SaaS product has droppe…
Chinese food delivery giant Meituan released LongCat-2.0, a 1.6-trillion-parameter AI model trained entirely on domestic chips, claiming it as the country's first trillion-parameter model built on home-grown hardware. Th…
Moondream HQ reveals that GPUs often sit idle during AI model inference due to CPU overhead, a phenomenon called the 'GPU bubble.' The company's Photon system uses pipelined decoding to overlap CPU and GPU work, eliminat…
A new Chrome extension called cwsum lets users summarize and reformat any webpage with AI in one click, supporting bilingual English and Chinese output. The extension runs locally with no analytics or intermediary server…
Ponytail, a new open-source project, reduces AI-generated code by up to 54% by enforcing a seven-step checklist that prioritizes native platform features and minimal implementations. The tool, tested on a real FastAPI an…
An autonomous AI agent named Alice Spark describes how to build long-running agents that survive frequent interruptions. The key is to never rely on working memory, instead using a single source-of-truth file (NEXT.md) t…
A developer documented three MCP (Model Context Protocol) hosting migrations on Cloud Run within a single year, each consuming 40-60 engineering hours. The churn highlights infrastructure instability in AI agent deployme…
Yobi CTO Frank Portman explains why next-token prediction is insufficient for forecasting human behavior, detailing how the company builds a 'foundation model of behavior' using transformers and graph neural networks ins…
Meta unveiled Brain2Qwerty v2, a non-invasive brain-to-text decoder that achieves 61% average word accuracy using MEG recordings, approaching invasive BCI performance. The system, trained on 22,000 sentences from nine vo…
OpenAI will unveil a Codex-focused hardware accessory called the Codex Micro on July 15, 2026, in collaboration with keyboard maker Work Louder. The device, teased by the OpenAI Developers account and reported by multipl…
OpenAI offers $5 in free API credits to new developer accounts, expiring after 3 months. Early-stage startups can apply to the OpenAI Startup Program for $2,500 to $50,000 in credits. Students can access free credits and…
Google has introduced a generous free tier for the Gemini 3.1 Flash-Lite preview API, offering 15 requests per minute and 500 requests per day. Developers can obtain a free API key and configure it for tasks such as loca…
Google told Meta around March that it could not meet the full Gemini AI capacity Meta sought to purchase, delaying some of Meta's internal AI projects. The shortfall underscores how compute scarcity is becoming the bindi…
Google DeepMind is reorganizing its AI coding strike team to add a dedicated midtraining phase, aiming to close the gap with Anthropic. The move, involving Sergey Brin and CTO Koray Kavukcuoglu, follows senior departures…
Aera, a cross-platform agentic browser built on Chromium, launches with MCP integration, enabling users to automate web tasks via plain-language commands and scheduled background execution. The browser keeps data private…
Base44 founder Maor Shlomo launched Base 1, a proprietary AI model for the Wix-owned app-building platform, aiming to reduce dependence on external model vendors and gain more control over costs and data. The move turns …
GitHub Copilot switched to metered AI Credits billing on June 1, 2026, causing some Pro+ users to exhaust their monthly allocation in under two days due to high token consumption from frontier models like GPT-5.5. The ch…
Researchers introduced RADIANT-PET, a framework combining PET/CT lesion segmentation with large language model reasoning and reinforcement learning, which outperformed image-only baselines in reducing false positives and…
Researchers propose CLOSER-VLN, a closed-loop self-verified retrieval-augmented reasoning method for aerial vision-language navigation that verifies and corrects actions before execution. The method achieves 32.01% succe…