Blog

OpenAI launches GPT-OSS 120B & 20B. First open-source weights since GPT-2. Benchmark details, pricing, and deployment guide.

Anthropic's latest reasoning model pushes boundaries with 200K context and advanced tool use for enterprise developers.

Anthropic launches Claude 4.5 Sonnet on July 29, 2025, claiming superior creative writing and multi-step reasoning capabilities for enterprise developers.

Zhipu AI releases GLM-4.5 Air, a lightweight 106B MoE variant optimized for efficient inference on H20 GPUs with open-source weights.

Zhipu AI releases GLM-4.5, a 355B parameter MoE model offering superior reasoning and coding capabilities at a fraction of the cost of competitors like DeepSeek.

xAI launches Grok 4 on July 11, 2025, marking a historic leap in reasoning capabilities and training infrastructure on the expanded Colossus cluster.

Google DeepMind releases Gemma 3n, a 4B parameter model optimized for mobile and edge devices under Apache 2.0, marking a new era of efficient on-device intelligence.

OpenAI releases GPT-o3 Pro, a non-open-source reasoning powerhouse designed for complex frontier problems with extended thinking capabilities.

Mistral AI releases Small 3.2 with improved instruction following and Apache 2.0 licensing for enterprise developers.