Anthropic: Claude Sonnet 4.5

Claude Sonnet 4.5 is a text model from Anthropic, released by the provider on 29 September 2025. It sits in Anthropic's Sonnet line, which the provider describes as its most advanced Sonnet model to date, optimized for real-world agents and coding workflows. On api14 it is listed under the model id anthropic/claude-sonnet-4.5 and falls into the text category.

The model accepts text, images and files as input and returns text only. Its context window is 1,000,000 tokens, and the maximum output is 64,000 tokens. The knowledge cutoff is 31 January 2025, so anything after that date is outside what the model was trained on. Because images and files are accepted alongside text, you can send mixed prompts that combine written instructions with visual or document content, while planning for text as the sole response format.

Those capabilities point to a few practical uses. Coding assistance and agent-style workflows are the areas Anthropic highlights, and the large context window suits tasks where you need to keep a long codebase, a lengthy document or a long conversation in view at once. Image and file input make it a reasonable fit for document analysis, extracting or summarizing information from uploaded material, and question answering over mixed content. Tool calling and structured outputs also make it usable as the reasoning step inside a larger pipeline, where the model decides which tools to invoke and returns data in a shape your code can parse.

You call it through api14's OpenAI-compatible API using the model id anthropic/claude-sonnet-4.5, so existing OpenAI SDK code works with a change of base URL and model name. The supported request parameters are include_reasoning, max_completion_tokens, max_tokens, reasoning, response_format, stop, structured_outputs, temperature, tool_choice, tools, top_k and top_p. In practice that means you can control sampling with temperature, top_p and top_k, cap output length with max_tokens or max_completion_tokens, define functions for the model to call with tools and tool_choice, and request structured or formatted responses with response_format and structured_outputs. Reasoning can be requested and its output included via the reasoning and include_reasoning parameters.

One limit worth noting is that output is text only, so if you need generated images or audio, this model is not the right choice. The knowledge cutoff of 31 January 2025 also means recent events are not covered unless you supply them in the prompt or through tools. If your workload is small and simple, a lighter model may be sufficient, but for long-context coding, document and agent tasks, the combination of a 1,000,000 token context window, 64,000 token output ceiling and tool support is the reason to pick this one.