OpenAI is reportedly preparing to launch a new AI model, possibly GPT-4.5 or GPT-5, in the coming weeks, with internal testing and API integrations already underway according to sources close to the company.
Anthropic is testing Claude Tasks mode, which separates planning from execution to function as a workflow runner with visible steps and user intervention.[5]
OpenAI shipped GPT Image 1.5 directly inside ChatGPT, improving text rendering, composition control, and iterative edits within a single thread.[5]
Google launched Gemini 3 Flash as a high-speed model, alongside Gemini 3 which delivers state-of-the-art multimodal reasoning integrated into Google Search and Android apps.[1][5]
OpenAI released GPT-5.2-Codex, a coding-specialized model update optimized for long context, multi-file edits, repo navigation, and tool-driven workflows in ChatGPT.[5]
Google confirmed Gemini will fully replace Google Assistant in 2026, with expansions to Wear OS, Google TV, Android Auto, and new features like Nano Banana for image editing.
GPT-5.2 Thinking mode achieved breakthrough 70.9% on GDPval, marking the first model to match human expert performance across 44 occupations.
Google's Gemini 3 Pro reached a historic 1501 Elo score on LMArena, the highest ranking ever, highlighting superior performance in late 2025 benchmarks.
Anthropic's Claude Opus 4.5, launched November 24, 2025, achieved a record 80.9% on SWE-bench Verified, leading in coding performance benchmarks.
OpenAI released ChatGPT's GPT-5.2 on December 11, 2025, introducing three operating modes: Instant for low latency, Thinking for deep reasoning, and Pro for maximum intelligence, with a 400K token context window and memory across sessions.
Anthropic launched Claude Opus 4.5 on November 24, 2025, achieving record 80.9% on SWE-bench, positioning it strongly for coding tasks.[2][3]