Anthropic's Claude Opus 4.5 Challenges Gemini, GPT

Anthropic has released Claude Opus 4.5, its new flagship model aimed squarely at Google’s Gemini 3 and OpenAI’s GPT-5.1. The announcement highlights Opus 4.5’s strength on coding and agentic benchmarks, positioning it as a contender for top performance in the current frontier-AI sprint.

According to the company coverage, Opus 4.5 is the first model to exceed 80% on the SWE‑Bench Verified coding benchmark and sets new highs in tool use, reasoning, and problem-solving. Anthropic says the model matches or beats Gemini 3 across multiple benchmarks and emphasizes Opus as “the most robustly aligned model” in safety terms. Architecturally, Opus is designed to orchestrate teams of smaller Haiku models, serving as a central coordinator in multi-agent setups. Anthropic also announced a notable 66% price cut from Opus 4.1 alongside claimed efficiency gains versus its prior models.

Alongside Opus 4.5, Anthropic rolled out product updates such as unlimited chat lengths, Claude Code on desktop, and broadened access to Claude integrations for Chrome and Excel. The timing—coming in the same week as Gemini 3 and GPT‑5.1 Pro—marks another rapid escalation in the frontier model race and underscores pricing and efficiency as key competitive levers.

Source