AI — Page 31

The Qwen3.5-based research preview includes 4B, 9B and 27B checkpoints, while withholding training data and restricting the code to noncommercial use.

The app was due August 1st; Google's existing product already lets builders create and distribute software through the browser.

Alexi Gladstone's open-source method spends extra compute during training and reports gains across image, video and masked language models.

The FutureHouse spinout is publishing weights, code and benchmarks to support its push to sell Kosmos into biopharma R&D.

Solo founder Adam Rida built the YC S26 system to allocate compute across open-weight models through one OpenAI-compatible endpoint.

The framework tests Claude Code, Codex and OpenCode on database, auth and Edge Function tasks, then publishes the scores.

Founder Trenton Hughes is selling recurring content and growth operations to B2B companies, with human experts supervising AI output.

Founders Laodis Menard and Brivael Le Pogam are placing MiniMax's video model inside Argil's production system for avatars, scripts, captions, B-roll and finished social clips.

The 512K-context release tests whether Huawei can turn its chip, model and inference stack into a credible alternative to Nvidia-centered AI.

The OpenAI CEO made the claim after the company reported a 20% serving-cost reduction and greater than 15% token-efficiency gain from GPT-5.6 Sol, while passing some savings to Luna and Terra users.

Miguel Acevedo's macOS app turns research, files, recurring tasks and outputs into an inspectable workspace.

DeepSeek's July 31st public beta adds native Responses API support while leaving V4-Pro and the models in its app and website unchanged.

Leopold Aschenbrenner kept a private Anthropic stake reportedly worth $5 billion, preserving his largest route back from the rout.

The U.S. rollout gives AI Pro and Ultra subscribers an agent that can use active browser sessions, with daily task caps and security warnings.

The July 30th announcement centers on a LoRA adapter first published in February and still awaiting an official DrugProt test-set evaluation.

The YC-backed founders say Deepgram, Giga and Mintlify use Cardinal to design and run sales workflows.

CTGT reports finance gains without the teacher's political refusals, though unreleased adapters prevent full independent reproduction.

Paid Nous Portal subscribers get temporary access to Black Forest Labs' audio, image and video model, plus a short-film contest.

Ellis raised an $11 million seed led by First Round in April, as the Cadre founder sells an AI operations layer to smaller credit managers.

The API-only model aims to remove brittle text-normalization rules, with 1 million credits offered for production-case feedback.

Alex Reibman's 24-hour experiment added five users and no revenue after Saul hit distribution blocks and bought a $99.50 user-testing campaign.

The proxy swaps placeholder keys for real credentials at the network boundary, extending Infisical's secrets platform into agent infrastructure.

The 276-billion-parameter multimodal model gives Murati's Tinker platform a cheaper base for fine-tuning coding agents and custom AI apps.

The reward-model system uses millions of user preferences and scored above 0.98 correlation with Arena's later live rankings.