Anthropic releases Claude 4.5, calls it their most advanced model yet

449

Anthropic has unveiled Claude 4.5, its most capable model to date, marking another leap in the increasingly competitive AI model race. Described as a “major upgrade,” Claude 4.5 improves performance across reasoning, coding, math, and multimodal tasks—and reportedly outperforms OpenAI’s GPT-4 and Google’s Gemini 1.5 Pro on several benchmarks.

According to Anthropic, Claude 4.5 scored better than GPT-4 in 10 out of 11 major industry tests, including HumanEval (coding), MMLU (language understanding), and GSM8K (grade-school math). It also offers faster outputs, fewer refusals, and a more “friendly” interaction style.

The Claude 4.5 release is part of Anthropic’s broader plan to iterate at a monthly cadence—signaling both product maturity and user demand. While the company hasn’t added image generation yet, Claude 4.5 does support image analysis, making it useful for real-world applications such as interpreting charts, photos, and screenshots.

Importantly, Anthropic’s team highlights that Claude 4.5 still leans on their signature Constitutional AI approach, which aims to deliver safer, values-aligned answers through built-in guidance principles. Claude 4.5 is now available to users via claude.ai, the iOS app, and business APIs through Amazon Bedrock and Google Cloud’s Vertex AI.

As AI development cycles continue to compress, Claude 4.5 reinforces how smaller teams like Anthropic are holding their own against the giants. With GenAI adoption growing across enterprises, such incremental model updates are increasingly critical for real-world performance, not just lab demos.