Latest news, tutorials, and insights about AI

Moonshot AI releases Kimi K2, a 1T parameter open model ranking #1 on LMSYS. See pricing, benchmarks, and specs for developers.

Sarvam AI releases Sarvam-2B, an open-source 2B parameter model optimized for 10+ Indian languages, marking a new step in sovereign AI infrastructure.

Korea's definitive answer to the open frontier model category. SOLAR 102B brings 128k context and MoE efficiency to the developer community.

Google DeepMind releases Gemini 3 Flash on December 17, 2025. Explore the architecture, pricing, and benchmarks of this new frontier-class model.

Allen AI releases Molmo 2, an 8B parameter multimodal model with full open weights, challenging industry standards in reasoning and visual understanding.

Xiaomi releases MiMo V2 Flash, a 309B MoE reasoning model optimized for math and code, challenging GPT-5 benchmarks at a fraction of the cost.

OpenAI unveils GPT-5.2, featuring expert-level performance on 44 tasks, 128K output tokens, and enhanced mental health safeguards for developers.

Mistral AI introduces Devstral Small 2, a successor to Devstral Small 1 derived from Mistral Small 3.1, featuring 24B parameters and Apache 2.0 licensing for portable coding agents.

Mistral AI releases Devstral 2, a 123B parameter open-source coding model with top-tier SWE-Bench performance and a revenue-based license.