Google DeepMind has launched Gemini 2.5 Deep Think, its most advanced AI reasoning model to date, now available exclusively to Google AI Ultra subscribers at $250 per month via the Gemini app. This model is a variation of the version that achieved a gold medal at the 2025 International Mathematical Olympiad (IMO) and employs parallel thinking and reinforcement learning to brainstorm multiple solutions simultaneously. It demonstrates strong performance across various benchmarks, including a 34.8% score on Humanity's Last Exam without external tools, 86.6% on Live Code Bench, and 99.2% on the AIME 2025 math competition. The model is designed to assist researchers, scientists, and academics in tackling complex scientific literature, coding problems, and strategic planning tasks. Meanwhile, Anthropic has released Claude Opus 4.1, an upgrade to its Claude Opus 4 model, enhancing capabilities in agentic tasks, real-world coding, and complex reasoning. Claude Opus 4.1 achieves a 74.5% score on the SWE-Bench, surpassing both Gemini 2.5 Pro and OpenAI's o3 model, and is currently available to paid users and via API. In a related development, Anthropic revoked OpenAI's access to its Claude models, citing violations of terms of service related to OpenAI's internal use of Claude Code ahead of the GPT-5 launch. OpenAI has responded by releasing two new open-weight models, GPT-OSS-120B and GPT-OSS-20B, marking its first open-source model release since GPT-2. These models are available under the Apache 2.0 license and can be run locally on consumer hardware, with the 120B parameter model performing near parity with OpenAI's o4-mini model and the 20B parameter model delivering results comparable to o3-mini. This release represents a strategic shift for OpenAI, emphasizing accessibility and developer control. The AI industry is witnessing rapid advancements and competitive releases across major players including Google DeepMind, Anthropic, and OpenAI, with significant updates expected in the coming weeks including OpenAI's anticipated GPT-5 launch.