Anthropic: Claude Opus 4.7 (batch)

Claude Opus 4.7 (batch) is a text generation model from Anthropic, part of the Opus family and positioned as the next generation of that line. It is built specifically for long-running, asynchronous agents, and it builds on the coding and agentic strengths of Opus 4.6. The batch variant is the version you call through api14, and the model id to use is "anthropic/claude-opus-4.7:batch".

The model accepts text, images and files as input, and produces text as output. Its context window is 1,000,000 tokens, which is large enough to hold very long documents, extended conversation histories or multi-file inputs in a single request. The maximum output length is 128,000 tokens. Because it handles images and files as well as text, you can pass mixed content directly rather than extracting everything to plain text first.

Those capabilities point to a few practical uses. Long document analysis and summarisation fit naturally, since a million-token window lets you work over large contracts, reports or codebases without chunking. Agentic and tool-driven workflows are another fit: the model supports tools and tool_choice, so you can build assistants that call functions, retrieve data and act over many steps. Coding tasks, such as reviewing or generating code across a repository, also follow from the combination of a large context and file input. Multimodal tasks like answering questions about charts, screenshots or scanned documents are possible too, given image input.

To call it through api14, use the OpenAI-compatible endpoint with the model id "anthropic/claude-opus-4.7:batch". The request supports include_reasoning, max_tokens, reasoning, reasoning_effort, response_format, stop, structured_outputs, tool_choice, tools and verbosity. In practice this means you can cap output with max_tokens, control reasoning depth with reasoning_effort or the reasoning parameter, request structured JSON via response_format or structured_outputs, and define callable functions with tools and tool_choice. Because the API is OpenAI-compatible, existing OpenAI client libraries work with only the base URL and model id changed.

The batch designation is the main thing to keep in mind when choosing this model. It suits workloads where you can tolerate asynchronous execution in exchange for the model's scale and agentic behaviour, such as bulk document processing or background agent runs. If you need synchronous, latency-sensitive responses, or your task is small and simple, a non-batch model may be the better fit. The facts here describe the model's inputs, outputs, context and parameters; the live price table on this page covers cost.