Claude Opus 4.6 and Sonnet 4.6 now offer the full 1M context window at standard pricing, eliminating the previous long-context premium. Standard pricing applies across the entire window, meaning a 900K-token request costs the same per-token rate as a 9K one. Media limits expand to 600 images or PDF pages, up from 100. This general availability arrives on the Claude Platform, Microsoft Azure Foundry, and Google Cloud’s Vertex AI, requiring no beta header or code changes for existing requests over 200K tokens.

This shift removes the need for lossy context summarization and compaction, a major pain point for users handling large datasets or codebases. Opus 4.6 achieves the highest MRCR v2 score (78.3%) among frontier models at this context length, enabling more accurate reasoning across vast information. Users can now load entire codebases, lengthy contracts, or full agent traces without losing fidelity. Claude Code users on Max, Team, and Enterprise plans default to 1M context automatically, significantly improving agent performance on tasks like code reviews and legal analysis.

Engineering teams report tangible benefits: a 15% reduction in compaction events and more efficient token usage. Large diffs that previously required chunking now fit seamlessly, allowing agents to maintain cross-file dependencies and deliver higher-quality outputs. The expanded context unlocks complex real-world applications, from legal document analysis to scientific research synthesis across multiple papers and codebases in a single pass.