1M Context Window Now Default for Claude Code on Max, Team, and Enterprise
Claude CodeAnthropic has made the full 1-million-token context window the default for Claude Code users on Max, Team, and Enterprise plans running Opus 4.6, eliminating the previous extra-usage charge and beta header requirement. The change — announced via the official Claude blog on March 13, 2026, alongside the general availability of 1M context across the Anthropic API — means long coding sessions will experience significantly fewer compaction events, keeping more of the active conversation intact. Standard per-token pricing ($5/$25 per million tokens for Opus 4.6) now applies uniformly across the entire window with no long-context multiplier. Media limits have also expanded from 100 to 600 images or PDF pages, benefiting workflows that involve large documentation sets.
Sources & Mentions
5 external resources covering this update
Claude Code Changelog: 2.1.75 release
X (Twitter)
Anthropics Claude Opus 4.6 brings 1M token context and agent teams to take on OpenAIs Codex
VentureBeat
Claude Opus 4.6 Introduces Adaptive Reasoning and Context Compaction for Long-Running Agents
InfoQ
Claudes 1M Context Window Is Now Generally Available
Dev.to
[AINews] Context Drought
Latent Space
1M Context Window Is Now the Default for Claude Code
Starting March 13, 2026, Claude Code users on Max, Team, and Enterprise plans no longer need to configure anything to access the full 1-million-token context window with Opus 4.6. Anthropic shipped the change as part of v2.1.75, simultaneously with the general availability announcement on the Claude blog — marking a meaningful upgrade for developers working with large codebases, complex agent traces, or extensive documentation.
What Changed in v2.1.75
Previously, accessing the full 1M context window in Claude Code required additional usage charges above standard plan limits. With v2.1.75, Opus 4.6 sessions on Max, Team, and Enterprise plans automatically use the full 1M-token window — no extra cost, no beta headers, no manual configuration.
The practical result is a significant reduction in compaction frequency. Claude Code auto-compaction kicks in when the context window fills up, summarizing earlier portions of the conversation to free space. With 1M tokens available by default, sessions can run far longer before compaction becomes necessary. Anthropic has reported a 15% decrease in compaction events since enabling 1M context support.
Pricing and API Changes
The broader 1M context GA applies the same standard per-token pricing across the entire window:
- Opus 4.6: $5 per million input tokens / $25 per million output tokens
- Sonnet 4.6: $3 per million input tokens / $15 per million output tokens
There is no multiplier for long requests. A 900,000-token request costs the same per token as a 9,000-token one. The anthropic-beta: long-context-2025-01-01 header that was previously required for requests over 200K tokens is no longer needed, though existing code using it continues to work.
Expanded Media Limits
Alongside the context expansion, the maximum number of images or PDF pages per request has grown from 100 to 600. For Claude Code workflows that involve large documentation sets, design specifications, or multi-file PDF analysis alongside code, this removes a previously limiting constraint.
Impact on Claude Code Workflows
The 1M context window is particularly valuable for:
- Large monorepos: Entire codebases can be loaded without requiring selective file pruning or frequent context clearing.
- Long agent sessions: Multi-step agentic tasks that generate large tool-call traces, observations, and intermediate reasoning can remain intact throughout the session.
- Documentation-heavy work: Technical documentation, API references, and specifications can be kept in context alongside code without displacement.
Claude Code compaction behavior remains available and continues to function as a safety net, but for Max, Team, and Enterprise users on Opus 4.6, it will now engage far less frequently by default.