Vigilant PR v1 is a portable, workflow-agnostic AI pull-request reviewer that posts review comments on behalf of the user's GitHub identity, not a generic bot. Developed by tllongdev, it supports any …
Chinese AI models Kimi K3 and DeepSeek are forcing Silicon Valley to rethink assumptions about the AI race, with Moonshot pausing new Kimi K3 subscriptions less than 48 hours after launch due to overw…
A June 2026 survey of 145 enterprises by VentureBeat found two-thirds had already adopted a diversified AI strategy before Anthropic's Fable 5 shutdown, with 90% of enterprise model spend going to thr…
Lynkr, an open-source LLM router, scored 67.65 on the RouterArena benchmark, landing 15th of 27 routers. The router achieved a cost of $0.29 per 1K queries, outperforming GPT-5 as a router at 34× lowe…
Experiential Labs released World-Model-Optimizer, a CLI tool that distills frontier large language models into specialized local agents by using a router to decide when to call a frontier model versus…
World-Model-Optimizer, a tool from Experiential Labs, cuts LLM costs by over 40% through distillation and routing, using agent traces to refine smaller models for repetitive tasks while reserving fron…
OpenRouter's Fusion compound API, which fans prompts to multiple models and combines answers with a judge model, beat most frontier models in tests, with a panel of cheap models landing within 1% of C…
A developer scrapped 9 out of 10 planned LLM evaluation experiments after running only CI diagnostics, finding that one real-world use case—triaging flaky integration tests across multiple languages—p…
Experiential Labs launched wmo serve, an open-source tool that routes repetitive agent tasks to distilled smaller models, achieving frontier quality at 40%+ lower cost. The tool uses an OpenAI-compati…
Maginary.ai, an image/video generator with a Midjourney-like prompt syntax using 40+ underlying models, announced a major update integrating seedance2 and GPT-images2 support. The platform, which acts…
Roo Code, an agentic AI coding tool forked from Roo Cline, outperforms Cursor in shipping working code by autonomously executing tasks, reading terminals, and iterating until tests pass, according to …
Thinking Machines Lab released Inkling on July 15, a 975-billion-parameter open-source mixture-of-experts model trained from scratch, making it the best open-source model from a Western lab. The model…
Open-weight AI models now handle roughly 70% of production tokens routed through OpenRouter, up from 30% a year ago, as enterprises shift from frontier commercial models to cheaper open alternatives. …
Chinese open-weight AI models are capturing significant market share in the US as developers shift workloads to cheaper alternatives, with Coinbase CEO Brian Armstrong reporting the company cut AI spe…
Scalattice, founded by a 2x founder, launched an OpenAI-compatible inference API that runs open models like Qwen3 and Llama 3.3 70B on a distributed network of consumer-hosted machines. Providers earn…
Hydra, a local-first trust control plane for AI, routes each task to the cheapest model that meets a user-defined confidence threshold, running fully offline with no daemon or cloud dependency. The to…
Empire LLM for Codex, a new plugin from Web5 Labs, enforces a strict boundary where external AI models provide evidence but not authority, keeping Codex as the sole decision-maker. The plugin routes b…
Oleksii Trekhleb built Yes-Brainer, a web app that sends one question to multiple large language models simultaneously and offers three deliberation modes — Parallel, Trial, and Consensus — to surface…
GitHub Models, the free AI inference playground launched in September 2024, will shut down permanently on July 30, with no grace period for users relying on its playground, model catalog, inference AP…
An engineer from the Agent K project reports that their incident-response agent correctly diagnosed root causes in only 5 out of 12 seeded production incidents, yet it never executed a bad rollback. T…