Overview
Anthropic shipped Claude Code v2.1.75 on its standard release cadence, expanding the default context window for Claude Opus 4.6 to 1 million tokens for Max, Team, and Enterprise plan subscribers. Previously, accessing 1M context on Opus 4.6 required additional usage or was gated separately from default behavior.
This is a feature-availability change, not a new model release. The underlying Opus 4.6 model is unchanged. What changed is that the 1M context window now activates by default rather than requiring explicit opt-in or extra usage. The same release also added session color customization, memory file timestamps, and several bug fixes, but the context window expansion is the change with direct implications for how the model processes and surfaces information.
What this means for brands
A 1M context window means Claude Code can now ingest significantly larger codebases, documentation sets, and knowledge bases in a single session without truncation or compaction interrupting the task. For brand operators building or maintaining AI-assisted workflows, this changes retrieval behavior: the model can hold more of your brand guidelines, product documentation, and competitive context in a single pass, reducing the likelihood that relevant brand information gets cut off mid-session. Outputs generated from long-context sessions are more likely to reflect the full scope of your provided materials rather than a truncated slice.
The practical second-order effect is on consistency. Shorter context windows force compaction, which introduces summarization artifacts. With 1M tokens available by default, sessions running on Max, Team, or Enterprise can process large brand asset libraries or lengthy style guides without the model losing early context. If your team uses Claude Code to generate marketing copy, documentation, or customer-facing content at scale, outputs should be more coherent across long sessions.
What to do
If your team is on Max, Team, or Enterprise, verify the upgrade applied by checking the session header in Claude Code after updating to v2.1.75 or later. The context indicator should reflect 1M rather than the prior default. From there, run a test session that loads your full brand guidelines, tone-of-voice documentation, and any product reference material you previously had to split across sessions. Note whether output quality or consistency differs compared to sessions that triggered compaction.
Audit any internal workflows that were structured around the old context limit. If you built chunking logic, document-splitting pipelines, or multi-session handoffs specifically to work around the previous window size, those may now be unnecessary overhead. Identify the three to five most context-heavy use cases your team runs on Opus 4.6 and retest them directly. If your plan tier is below Max or you are on API-key access without a qualifying subscription, the 1M default does not apply and your existing setup is unchanged.
