Unofficial page — not affiliated with or endorsed by Anthropic. Chat answers are AI-generated and may be wrong.
Anthropic · Anthropic’s own facts
Claude Fable 5.2 vs Claude Opus 5.5 (2026)
Claude Fable 5.2 costs twice as much as Claude Opus 5.5 ($8 / $40 vs $4 / $20 per million input / output tokens), both have a 1M-token context and 128K-token output, and Anthropic recommends starting most work with Opus 5.5.
Per million tokens
Price per million tokens
| - | Input | Output | OpenRouter id |
|---|---|---|---|
| Claude Fable 5.2 | $8 | $40 | anthropic/claude-fable-5.2 |
| Claude Opus 5.5 | $4 | $20 | anthropic/claude-opus-5.5 |
Anthropic’s own facts
Side by side
| - | Claude Fable 5.2 | Claude Opus 5.5 |
|---|---|---|
| Released | November 9, 2026Source: platform.claude.com · checked 2026-11-11 | September 22, 2026Source: anthropic.com · checked 2026-11-11 |
| Price per 1M tokens (input / output) | $8 / $40Source: platform.claude.com · checked 2026-11-11 | $4 / $20Source: anthropic.com · checked 2026-11-11 |
| Fast mode (input / output) | $16 / $80Source: platform.claude.com · checked 2026-11-11 | $8 / $40Source: anthropic.com · checked 2026-11-11 |
| Context / max output | 1M / 128K tokensSource: platform.claude.com · checked 2026-11-11 | 1M / 128K tokensSource: anthropic.com · checked 2026-11-11 |
| Default effort / latency | high / ModerateSource: platform.claude.com · checked 2026-11-11 | medium / ModerateSource: anthropic.com · checked 2026-11-11 |
| Input → output | Text and image → textSource: platform.claude.com · checked 2026-11-11 | Text and image → textSource: platform.claude.com · checked 2026-11-11 |
| Knowledge cutoff | June 2026Source: platform.claude.com · checked 2026-11-11 | June 2026Source: anthropic.com · checked 2026-11-11 |
| Retirement (not sooner than) | November 9, 2027Source: platform.claude.com · checked 2026-11-11 | September 22, 2027Source: anthropic.com · checked 2026-11-11 |
| API id | claude-fable-5-2Source: platform.claude.com · checked 2026-11-11 | claude-opus-5-5Source: anthropic.com · checked 2026-11-11 |
| What Anthropic says it is for | For demanding reasoning and long-horizon agentic workSource: platform.claude.com · checked 2026-11-11 | For long-running agentic coding and knowledge workSource: platform.claude.com · checked 2026-11-11 |
Which one does the maker recommend?
- Anthropic recommends starting most work with Claude Opus 5.5, and using Claude Fable 5.2 for more demanding reasoning and long-horizon agent work, or when your evaluations of Opus 5.5 at higher effort still fall short.Source: platform.claude.com · checked 2026-11-11
- In AI research and development work such as accelerator kernel optimisation, Anthropic’s launch post puts Fable 5.2 at about the Opus 5.5 level.Source: anthropic.com · checked 2026-11-11
Sample answers (AI-generated)
Same question, both models
Each model was called once with the same question on 2026-11-11; answers are shown exactly as returned (plain text), not edited.
Question: In two sentences, what is a context window in a language model?
Claude Fable 5.2 · AI-generated · anthropic/claude-fable-5.2
A context window is the maximum amount of text, measured in tokens, that a language model can consider at one time, including both the input it receives and the output it generates. Anything beyond that limit is effectively invisible to the model, so it cannot reference or remember information that falls outside the window.
Claude Opus 5.5 · AI-generated · anthropic/claude-opus-5.5
A context window is the maximum amount of text, measured in tokens, that a language model can consider at once, including both the input prompt and the response it generates. Anything beyond this limit falls outside the window, so the model cannot see or use it when producing its output.
Question: Give three short tips for writing a clear email subject line.
Claude Fable 5.2 · AI-generated · anthropic/claude-fable-5.2
1. Lead with the main point: put the key topic or request in the first few words so it survives truncation on mobile. 2. Be specific: write "Budget approval needed by Friday" instead of "Quick question." 3. Keep it short: aim for about six to eight words, and cut filler like "Hi" or "Just checking in."
Claude Opus 5.5 · AI-generated · anthropic/claude-opus-5.5
1. Be specific: state the exact topic or request, such as "Budget approval needed for Q3 campaign." 2. Keep it short: aim for six to ten words so it displays fully on phones. 3. Front-load key details: put deadlines, actions, or names first, like "Due Friday: Final report edits."
Question: In one short paragraph, what is the difference between weather and climate?
Claude Fable 5.2 · AI-generated · anthropic/claude-fable-5.2
Weather is the short-term state of the atmosphere in a specific place, such as today's temperature, rain, wind, or humidity, and it can change within hours or days. Climate is the average pattern of weather in a region measured over long periods, typically 30 years or more. In short, weather is what you experience on a given day, while climate is what you can generally expect based on long-term trends.
Claude Opus 5.5 · AI-generated · anthropic/claude-opus-5.5
Weather is the day-to-day state of the atmosphere in a specific place, including temperature, rain, wind, and humidity, and it can change within hours. Climate is the average pattern of weather in a region over long periods, typically 30 years or more. In short, weather tells you what to wear today, while climate tells you what kinds of clothes to keep in your closet.
Use it now
Chat
The chat starts with Claude Fable 5.2; switch to Claude Opus 5.5 in the model list to ask it the same thing.
FAQ
Questions
Which is cheaper, Claude Fable 5.2 or Claude Opus 5.5?
Claude Opus 5.5: $4 / $20 per million input / output tokens, half of Fable 5.2’s $8 / $40.
Which should I start with?
Anthropic recommends Claude Opus 5.5 for most work, and Fable 5.2 when you need more demanding reasoning or long-horizon agent work.
Do they have the same context window?
Yes. Both have a 1M-token context window and up to 128K output tokens, and both take text and images.
Which one is newer?
Claude Fable 5.2 (November 9, 2026). Claude Opus 5.5 came out on September 22, 2026.
Why are there no benchmark scores here?
We only list facts from Anthropic’s own pages that we have read. Scores we saw were quoted second-hand, so we left them out.
Are the sample answers official?
No. They are AI-generated answers we got by calling each model once with the same short question; they show style, not quality.