
Google Launches Gemini 3.1 Pro With Three Thinking Levels
Gemini 3.1 Pro enters preview in February 2026 with adjustable thinking modes, a verified 77.1% ARC-AGI-2 score, and expanded rollout across Gemini and developer tools.
The AI treasure trove. We scour all AI sources so you don’t have to. Find a compilation of valuable resources, libraries, and the most useful dictionary.

Gemini 3.1 Pro enters preview in February 2026 with adjustable thinking modes, a verified 77.1% ARC-AGI-2 score, and expanded rollout across Gemini and developer tools.

Claude Sonnet 4.6 narrows the gap with Opus 4.6, offering 1M token context, improved coding, and stronger office automation at lower cost.

ByteDance introduces Safeguards to Seedance 2.0 following legal threats from Disney and Paramount and warnings from the Motion Picture Association and SAG-AFTRA.

Alibaba introduced Qwen3.5 on February 16, 2026, pairing open-weight release with a hosted 1M-context model and efficiency gains designed for AI agents.

ByteDance released Seedance 2.0 in February 2026, enabling multimodal AI video generation with audio. Hollywood groups including the Motion Picture Association and SAG-AFTRA condemned the model over copyright concerns.

MiniMax unveiled M2.5 and M2.5-Lightning AI models that approach frontier coding performance while sharply lowering costs, with OpenHands adding external support for long-run agent use.

Anthropic released Claude Opus 4.6 with 1M-token context in beta, longer agent work, and stronger outputs for coding and office tasks.

OpenAI has released GPT-5.3-Codex and the Frontier orchestration platform. These tools enable enterprises to manage autonomous AI coworkers at scale.

Roblox 4D creation is now in open beta, letting creators and players generate functional, physics-driven 3D objects with Cube-powered schemas.