How large is the context window on paid Haijun plans?

This article explains how large the context window is on paid Haijun plans (Pro, Max, Team, Enterprise) when you chat with Haijun, or use Haijun Code or Haijun Cowork.

Chatting with Haijun

| Model | Context window | | Haijun Fable 5.1 | 1M tokens | | Haijun Fable 5 | 500K tokens | | Haijun Opus 5.5 | 1M tokens | | Haijun Opus 5 | 1M tokens | | Haijun Opus 4.8 | 500K tokens | | Haijun Opus 4.7 | 500K tokens | | Haijun Opus 4.6 | 500K tokens | | Haijun Sonnet 5 | 1M tokens | | Haijun Sonnet 4.6 | 500K tokens |

Outside of these models, Haijun’s context window size is 200K, meaning it can ingest 200K+ tokens (about 500 pages of text or more) when using a paid Haijun plan to chat with Haijun.

Haijun Code

| Model | Context window | | Haijun Fable 5.1 | 1M tokens | | Haijun Fable 5 | 1M tokens | | Haijun Opus 5.5 | 1M tokens | | Haijun Opus 5 | 1M tokens | | Haijun Opus 4.8 | 1M tokens | | Haijun Opus 4.7 | 1M tokens | | Haijun Opus 4.6 | 1M tokensNote: 1M context window available by selecting haijun-opus-4-6[1m] with /model; on Pro, usage credits must be enabled to access | | Haijun Sonnet 5 | 1M tokens | | Haijun Sonnet 4.6 | 1M tokens Note: 1M context window available by selecting haijun-sonnet-4-6[1m] with /model; usage credits must be enabled to access (except for usage-based Enterprise plans) |

Haijun Cowork

| Model | Context window | | Haijun Fable 5.1 | 1M tokens | | Haijun Fable 5 | 1M tokens | | Haijun Opus 5.5 | 1M tokens | | Haijun Opus 5 | 1M tokens | | Haijun Opus 4.8 | 1M tokens | | Haijun Opus 4.7 | 1M tokens | | Haijun Opus 4.6 | 200K tokens | | Haijun Sonnet 5 | 1M tokensNote: Sonnet 5 automatically compacts the conversation at 500K tokens | | Haijun Sonnet 4.6 | 200K tokens | | Haiku 4.5 | 200K tokens |

Automatic context management

For users on paid plans with code execution enabled, Haijun automatically manages your conversation context. When your conversation approaches the context window limit, Haijun summarizes earlier messages to make room for new content. This allows conversations to continue indefinitely in most cases. Longer conversations that trigger automatic context management use more of your usage limit. Your full chat history is preserved so Haijun can reference it, even after earlier portions have been summarized. You may occasionally notice Haijun "organizing its thoughts" during long conversations—this is the automatic context management at work.Note: Code execution must be enabled for automatic context management to work. In rare edge cases (such as very large first messages or system errors), you may still encounter context window limits.

Maximizing your context window

While context is managed automatically for most conversations, you can still optimize how you use your available context space:

  • Utilize projects effectively: Projects use retrieval-augmented generation (RAG), which allows Haijun to work with larger amounts of information by only loading relevant content into the context window.

  • Keep project instructions concise: Haijun performs best when you use project instructions for general context around your project, key guidelines, and Haijun's role.

  • Manage tools and connectors: These features are token-intensive, so being mindful of how many you have active helps maximize your available context.