Teamday

Chat & Reasoning Model · Anthropic

Claude Opus 5.5

Claude Opus 5.5 is Anthropic's newest Opus model, released September 22, 2026, and the model Anthropic tells most teams to start with. It succeeds Claude Opus 5 at a lower price ($4 / $20 vs $5 / $25 per 1M tokens), with a 1M-token context window and up to 128K output tokens. It is TeamDay's recommended model on the Claude Code harness.

Use Claude Opus 5.5 in TeamDay →All models

At a glance (verified 2026-09-23)

Vendor
Anthropic
Type
Chat & Reasoning Model
Context window
1M tokens
Pricing
Base: $4 in / $0.20 cache hits / $20 out per 1M tokens. Cache writes: $5 5m / $8 1h.

Best for

Complex codingLong-running agentsKnowledge workMulti-step reasoning

How to use Claude Opus 5.5 in TeamDay

Claude Opus 5.5 is the recommended default in TeamDay's chat model picker on the Claude Code harness. Agents and chats previously set to Opus 5 or Opus 4.x run on it automatically.

Harness: claude-code →Max →James →

Frequently asked questions

How much does Claude Opus 5.5 cost per million tokens?

Base pricing is $4 per 1M input tokens, $0.20 for cache hits, and $20 per 1M output tokens; cache writes cost $5 (5-minute) or $8 (1-hour). That is 20% below Claude Opus 5 on input and output, with cache hits 60% cheaper.

When should I pick Claude Opus 5.5 over Fable 5.1?

Start with Opus 5.5 for most work — it costs 2.5x less than Fable 5.1 ($4 vs $10 per 1M input tokens). Move to Claude Fable 5.1 for the hardest, longest-running tasks, or when Opus 5.5 at higher effort still falls short.

What changed from Claude Opus 5?

Lower prices, and thinking is now always on — reasoning effort is the only control, and Anthropic's API default drops to medium (TeamDay defaults to high). Forced tool use now returns an error, as on Claude Fable 5.1.

Is Claude Opus 5.5 available in TeamDay?

Yes — it is the recommended model on TeamDay's Claude Code harness, with reasoning effort selectable from low to max.

Related

Claude Opus 5.5is one of the models TeamDay's AI employees already run — no API keys, no infrastructure.

Hire your first AI employee →