Feature request: Offer a mid-tier context window (e.g. 400K) at standard Max rates

Feature Request

As a Max subscriber, I'd like a context window option between the default (~200K) and the full 1M — something like 400K — that is billed at standard rates rather than the 2x/1.5x premium that kicks in above 200K.

Problem

The current options are:

1. Standard context (~200K) — billed at normal Max rates
2. 1M context — billed at 2x input / 1.5x output rates, draining quota significantly faster

There's nothing in between. For many real-world coding sessions — especially in large codebases with extensive project guidelines, MCP server context, and multi-file edits — 200K fills up quickly and context compression kicks in, losing valuable earlier context. But jumping to 1M at premium rates feels excessive when all you need is a moderate increase.

Proposal

Offer an intermediate context tier (e.g. 400K tokens) that:

• Provides meaningfully more room than the current default
• Is billed at standard Max subscription rates (no premium multiplier)
• Doesn't require opting into the full 1M window and its associated cost

Even a modest increase to 400K would significantly improve the experience for sessions that regularly bump up against the 200K ceiling without needing anywhere near 1M.

Why This Matters

Claude Code sessions with large codebases routinely load substantial context: project guidelines (CLAUDE.md), MCP server instructions, file reads, tool results, and conversation history. The 200K limit means context compression happens mid-session, losing earlier file reads and reasoning. A 400K window at standard rates would let most sessions complete without compression while keeping costs predictable for Max subscribers.