Anthropic claims Claude Opus 4.5 has better vision, reasoning, and mathematics skills than its predecessors. Photo: Anthropic
Anthropic announced its latest AI model, Claude Opus 4.5.
The AI startup claims that Opus 4.5 is “the best model in the world for coding, agents, and computer use”. It also said that Opus 4.5 is “meaningfully better at everyday tasks like deep research and working with slides and spreadsheets.”
To support these claims, Anthropic pointed to the results across several industry benchmarks. The SWE-bench Verified benchmark, which is used to measure an AI system’s software coding capabilities, reveals that Opus 4.5 outperformed Google’s Gemini 3 Pro and OpenAI’s ChatGPT-5.1.
Other benchmarks, including Terminal-bench 2.0, 𝜏²-bench, MCP Atlas, OSWorld, and ARC-AGI-2 (Verified), also show Opus 4.5 leading its competitors across coding, reasoning, and agentic tasks.
Early testers of Opus 4.5 noted that it handled ambiguity and reasons about trade-offs without hand-holding. When presented with a complex, multi-system bug, Opus 4.5 was able to figure out a fix.
Claude says Opus 4.5 is now available on its desktop app, through its API, and across all three major cloud platforms.
Founded in 2021 by former OpenAI researchers and employees, Anthropic recently attracted billions of dollars in investment from Microsoft and NVIDIA. Opus is its largest and most capable model, followed by the mid-range Sonnet, and the lightweight Haiku models.
