Claude Opus 4.8

Claude Opus 4.8 adds seven capabilities but lacks performance benchmarks

Claude Opus 4.8 introduces seven new capabilities: tool use, vision, JSON mode, PDF input, reasoning, JSON schema, and prompt caching. These additions significantly expand the model's functionality, bringing it closer to feature parity with competing frontier models. The vision capability enables image understanding, while tool use allows function calling for extended workflows. PDF input support streamlines document processing, and the reasoning feature suggests enhanced problem-solving capabilities. JSON mode and schema support improve structured output generation, while prompt caching can reduce latency and costs for repeated interactions. However, this release includes no benchmark performance data across standard evaluation suites. Without metrics on coding ability, mathematical reasoning, instruction following, or general knowledge tasks, users cannot assess whether core model quality has improved, regressed, or remained stable. The absence of comparative scores makes it impossible to position this release against other frontier models or previous versions. Organizations considering adoption should conduct their own evaluations on task-specific benchmarks relevant to their use cases before deployment.