ThinkingBox trains AI agents using database rewards
Hugging Face: Turning Database Grading into RL Rewards: Post-Training Agents with ThinkingBox
Using database performance metrics as training signals could help AI systems learn to reason more effectively through reinforcement learning approaches.
Get smart on it
ThinkingBox is a system that trains AI agents by converting database verification checks into training rewards. The system evaluates whether an agent's actions produce the correct final database state, then uses those same executable checks to reward the agent during training on new tasks. When applied to a model over 50 training steps, this approach increased the share of tasks completed correctly on all attempts from 38 tasks to 107 tasks, while average success rose from 51.70% to 60.78%, demonstrating that agents can learn more reliable behavior when trained with rewards tied to actual database outcomes rather than just tool-call sequences.
Relevant courses
Want to go deeper than the news? Explore live, cohort-based AI courses taught by practitioners.
Browse AI courses on MavenRelated news
Canva releases its first image generation model
Canva is making its new image model available to all users, promising better, more customizable performance at lower costs.
Also reported by The Verge
Mistral AI launches Voxtral 3, a full voice agent stack that listens, thinks, and talks in real time
Mistral AI launched Voxtral 3, a new family of language models built for real-time voice agents. The system combines speech-to-text, text-to-speech, and a voice agent SDK, supports more than 50 languages, and lets agents tap into external tools during live conversations. It can run on your own infrastructure or through EU-hosted APIs to keep data under your control. The article Mistral AI launches Voxtral 3, a full voice agent stack that listens, thinks, and talks in real time appeared
Also reported by MarkTechPost
Ai2 releases OlmoEarth v1.3 state-of-the-art AI-ready satellite embeddings for the whole planet
Ai2, the Allen Institute for Artificial Intelligence, today released a satellite map embedding model called OlmoEarth v1.3 and precomputed annual embeddings for the whole planet at various resolutions with a browser-based map-making tool. The model compresses a year of satellite observations into a numerical fingerprint for each 10-by-10-meter patch of ground at the smallest resolution, […] The post Ai2 releases OlmoEarth v1.3 state-of-the-art AI-ready satellite embeddings for the whole pl
Also reported by Hugging Face
Recursive launches Flint, a laptop coding model it says was largely built by AI
Startup Recursive has released Flint, a free, laptop-sized coding model built largely by its own AI research system, as an early test of AI that improves itself.
Also reported by Wired AI and MarkTechPost
OpenAI makes its Jev competitor widely available as decision model competition heats up
OpenAI is launching its decision model to the world. Can it compete with the original Jev?
Build more responsive apps with the Decisions API
Fast, focused, low-cost decisions for any application.