Skip to main content

Claude Opus 4.6

Anthropic's Opus-tier chat model — the screenplay-writing workhorse of NGMC's chat nodes. Long-context, vision-capable, streamed token-by-token, with adjustable thinking effort.

Capabilities​

FeatureSupport
Chat (text-to-text)Yes, streaming
Vision (image understanding)Yes — up to 10 connected images
Document referencesYes — up to 100 text/document inputs
Context Window1,000,000 tokens
Max Output128,000 tokens
Reasoning EffortNone / Low / Medium / High (adaptive thinking)

Chat Nodes on the Canvas​

Add an AI Chat node and pick Claude Opus 4.6 in the inspector. Everything connected upstream becomes context:

  • Images — the model sees them (shot references, storyboards, style frames) and can critique or describe them.
  • Text nodes and documents — scripts, notes, and prior chat outputs feed in as source material.

The conversation lives on the node — iterate in place, and downstream nodes can reference the latest reply.

Reasoning Effort​

  • None / Low — snappy conversational replies, rewrites, brainstorms.
  • Medium / High — the model thinks before answering; worth it for structural work (beat sheets, plot-hole analysis, continuity checks across a long script).

What It's Best At​

  • Screenwriting — dialogue with distinct character voices, scene description in proper format, coverage-style feedback.
  • Long-document work — the 1M-token window holds a full feature script plus revisions plus your notes simultaneously.
  • Visual analysis — feed generated frames back in and ask what's off before re-rendering.

Limitations​

  • Text out only — it plans images/video but doesn't render them (chain to an image/video node)
  • No audio understanding
  • Effort tops out at High (X-High is a Fable 5 capability)

See Also​