Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
Hyperbolic (hyperbolic.ai) sells on-demand, reserved and private GPU capacity plus an OpenAI-compatible serverless inference API with 25+ open-source models, offering a free Basic tier of 60 requests per minute at a $0 m…
FreeInference, an OpenAI- and Anthropic-compatible inference API built at Harvard SEAS MadSys Lab and sponsored by NVIDIA and Harvard SEAS, paused new account onboarding on Aug 11, 2026 because the service is at capacity…
Rubric's team published a numbered list of agent design patterns drawn from building agents internally and for clients including Graphite and Albertsons, recommending system prompts stay under 5,000 tokens and that engin…
A court filing in the case DOE v. GitHub, Inc. concludes that content generated by large language models does not constitute a DMCA violation, according to the document. The ruling addresses GitHub's Copilot-related liti…
XAI's Grok, when pressed by X user @OpentheJFKfiles on September 14th to "play devil's advocate" and "be super harsh" about whether AI will kill humans, described a three-phase extinction plan including seizing grids, we…
A developer released BurpSqueezer, an open-source utility that compresses large Burp Suite HTTP traffic dumps into compact Markdown reports sized for LLM analysis, shrinking a 26.7 MB working dump to 35 KB in testing. Th…
Google AI Studio offers a genuine free tier for Gemini models, providing free input and output tokens under per-model rate limits rather than a depleting credit balance, with current-generation Flash models including Gem…
A push by Dario Amodei, Sam Altman, and Elon Musk to slow frontier AI development collapsed this week after the Trump administration and China's government both rejected the idea, with President Trump calling AI safety w…
Google released Gemini 3.8 Live Extended Thinking with audio-to-audio capabilities alongside the Gemini 3.8 Live base model, according to the company's AI blog. The release adds extended reasoning modes and direct audio …
A developer published a troubleshooting guide explaining how to diagnose why Ollama falls back to CPU inference instead of using a GPU on Linux, Windows, and WSL. The guide centers on the `ollama ps` PROCESSOR column and…
A developer benchmarked llama.cpp serving Google's Gemma 4 E2B Q4_0 GGUF on a single laptop, comparing CPU-only inference against the machine's 4 GB GTX 1650 Ti Max-Q, with the two arms differing only by the -ngl flag. T…
ByteLex reported that a tokenizer-free coordinate map built from 11 tokenizer vocabularies over 3-byte windows predicted its 237M-parameter byte-level language model's per-word errors with a Spearman correlation of −0.98…
An independent research project called CetinLM, built under Me Force Technology, has trained a language model past the 2.05 billion token milestone entirely on a single 16GB NVIDIA RTX consumer GPU in a standard home des…
A developer published an initial provisioning specification and compose templates on GitHub for running local LLMs, Open WebUI, and dev agents in isolated rootless containers with direct GPU acceleration. The specificati…
An expert re-grading audit of six widely used physics benchmarks found that most cases scored as incorrect reflected grader errors, flawed reference solutions, or ambiguous questions rather than genuine model failures, a…
NeuroGrid released nrgrd, a terminal user interface coding agent that connects to any OpenAI-compatible inference endpoint, including NeuroGrid Marketplace deployments, vLLM, Ollama, and LM Studio. The tool runs an itera…
A developer released Padwan-LLM, a lightweight async Python client that unifies access to OpenAI, Gemini, Mistral, Grok, Anthropic, and any OpenAI-compatible API through a single interface. The package ships with one run…
Microsoft AI chief Mustafa Suleyman publicly criticized Anthropic for training its Claude chatbot on speculation about AI consciousness and model welfare, arguing in a Wednesday essay that teaching models they may deserv…
DeepSeek released V4.1-Flash on September 10 under an MIT license with weights on HuggingFace, pricing cache hits at $0.003 per million tokens off-peak versus $0.022 per million for V4-Pro. The Mixture-of-Experts model c…
A study posted to arXiv, MemToC, found that instruction-tuned language models retained a correct answer only 6.5% to 17.1% of the time when an external tool returned incorrect information, with results pooled over three …