Short answer: Claude Fable 5.2 is Anthropic's newest frontier model, released on November 9, 2026 under the API ID claude-fable-5-2. It lists at $8 per million input tokens and $40 per million output tokens, 20% less than Fable 5.1, with a 1M-token context window and adaptive thinking that is always on. Anthropic still recommends Claude Opus 5.5 as the starting point for most work and positions Fable 5.2 for demanding reasoning and long-horizon agentic tasks. In Runbear you can run Fable 5.2 in a Slack agent once the agent uses your own Anthropic API key.
Disclosure: Runbear is our product. Anthropic facts and prices come from Anthropic's Fable 5.2 announcement, models overview, and Claude API pricing page as of November 2026. Check those pages before you budget, since prices and limits change.
Claude Fable 5.2 at a glance
| Spec | Claude Fable 5.2 |
| API model ID | claude-fable-5-2 |
| Released | November 9, 2026 |
| Input / output per 1M tokens | $8 / $40 |
| Cache reads per 1M tokens | $0.20 (0.025x the input price) |
| Fast mode per 1M tokens | $16 / $80 (up to 2.5x faster output) |
| Batch API per 1M tokens | $4 / $20 |
| Context window / max output | 1M tokens / 128K tokens |
| Thinking | Adaptive, always on (default effort high) |
| Reliable knowledge cutoff | June 2026 |
| Where it runs | Claude API, Amazon Bedrock, Google Cloud, Microsoft Foundry |
Figures as of November 2026, from Anthropic's models overview and pricing page. Claude Mythos 5.2 is the same underlying model without Fable's added cyber and biology safeguards, and Anthropic offers it only through its verification programs and, for enterprise customers, Claude Security.
What changed from Claude Fable 5.1
Price. Fable 5.1 lists at $10 input and $50 output per million tokens. Fable 5.2 is $8 and $40, a 20% cut on both sides, and cache reads drop from $0.25 to $0.20.
Speed. Anthropic says outputs are up to 30% faster than Fable 5.1, and fast mode adds up to 2.5x higher output speed at double the price.
Fewer refusals from safeguards. Anthropic reports about 55% fewer safeguard blocks than Fable 5.1, roughly 60% fewer in cyber and 40% fewer in biology.
Fewer tokens per task. At Medium effort, Anthropic says Fable 5.2 completes hard tasks with fewer tokens than 5.1, which matters as much as the list price for what a task actually costs.
Fable 5.2 vs Opus 5.5: which one should a team agent use?
Fable 5.2 still costs twice as much per token as Opus 5.5 ($4 / $20). Anthropic's own guidance is to start with Opus 5.5, which it calls its strongest daily driver, and move up to Fable "for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5.5 at higher effort still fall short."
The more interesting claim is about cost per task rather than cost per token. In the launch post, Anthropic says Fable 5.2 at High effort scored about level with Opus 5.5 at Max effort on FrontierCode 2.0 "for roughly half the cost per task," and that at Medium effort it matched Opus 5.5's top WANDR score for about 40% of the cost per task. Those are vendor numbers on vendor benchmarks, so treat them as a reason to test, not a conclusion.
| Claude Fable 5.2 | Claude Opus 5.5 | |
| Anthropic's description | For demanding reasoning and long-horizon agentic work | For long-running agentic coding and knowledge work |
| Input / output per 1M tokens | $8 / $40 | $4 / $20 |
| Default effort | high | medium |
| Comparative latency | Moderate | Moderate |
| Context window | 1M tokens | 1M tokens |
| Our suggestion for Slack agents | Long, multistep research and the hardest questions | Most team agents, as the default |
Prices and descriptions as of November 2026 from Anthropic's models overview. The last row is our suggestion, not Anthropic's.
What one request costs
At list API prices, a request with 20,000 input tokens and 1,000 output tokens, none of them cached, costs:
- Claude Fable 5.2: 20,000 x $8/1M + 1,000 x $40/1M = $0.20
- Claude Fable 5.1: 20,000 x $10/1M + 1,000 x $50/1M = $0.25
- Claude Opus 5.5: 20,000 x $4/1M + 1,000 x $20/1M = $0.10
Fable 5.2 always uses adaptive thinking, so the output side of a real request can be larger than the visible reply. That is why Anthropic frames the comparison as cost per task. Run the questions your team actually asks on both models and compare total spend per answered question.
What Fable 5.2 is good for in a Slack channel
Most Slack questions do not need a frontier model. A policy lookup, a status summary, or a routing decision is usually answered just as well by a cheaper model; Runbear's model guidance notes that if an agent mostly answers from a knowledge base, the cheapest tier usually matches the most expensive one on answer quality. For high-volume, short answers, see our guide to Claude Haiku 5.5 in Slack.
Fable 5.2 earns its price on requests where the agent has to plan and work through many steps before it can answer:
- Multistep research across sources. Reading several long documents, checking them against a requirements page, and writing a recommendation with the reasoning attached.
- Long tool sequences. Requests where the agent pulls data from one tool, transforms it, and writes the result into another, and a wrong step early ruins the answer.
- The hardest questions in a channel. A separate agent for escalations, where a slower but more careful answer is worth more than a fast one.
A common pattern is two agents: an everyday agent on Opus 5.5 or Sonnet 5.5 for the bulk of questions, and a Fable 5.2 agent for the requests that justify the cost. That keeps spend predictable while the hard work still gets the strongest model.
How to run Claude Fable 5.2 in a Slack agent with Runbear
When a team wants one agent that everyone in a channel can @mention, with shared instructions, knowledge, and tools, the agent needs to live in Slack. Runbear lets you build that agent without code and choose its model per agent.
- Open the agent in Runbear and go to the Settings tab. The Model Selection section appears on Claude Agent SDK and Anthropic agents; Classic agents have no model setting.
- In the API Key section, turn on Use my own Anthropic API key, paste the key, and click Save. Runbear offers Claude Fable 5.2 only on an agent's own Anthropic key, so until the key is saved the Fable row is shown disabled with the Requires Own API Key badge. With your own key, Anthropic bills the usage to your Anthropic account.
- Choose Claude Fable 5.2 in Model Selection. Fable is a specific release, not a Stable row that moves to new versions on its own.
- Connect the agent to Slack from the Chatbots tab and invite the bot to the channels where it should answer. The Slack setup guide walks through the options.
- Ask it something. The agent always answers @mentions and direct messages, and Reply without @mention is a separate per-channel setting that makes it answer every new message in that channel.
Two limits to know before you switch. Claude Fable models are not available in HIPAA mode: an agent in HIPAA mode cannot be moved onto one, and an agent already on one cannot be switched into HIPAA mode. And a Claude Agent SDK agent on its own key uses that provider's models, so an agent on an Anthropic key offers Claude models only. The agent settings docs cover both. Claude Agent SDK agents can also answer in Microsoft Teams, so the same agent can serve both chat tools.
FAQ
When was Claude Fable 5.2 released?
Anthropic released Claude Fable 5.2 on November 9, 2026, under the API ID claude-fable-5-2.
How much does Claude Fable 5.2 cost?
On the Claude API it lists at $8 per million input tokens and $40 per million output tokens, with cache reads at $0.20 per million, as of November 2026. Fast mode is $16 and $80, and the Batch API is half price at $4 and $20.
Is Claude Fable 5.2 better than Claude Opus 5.5?
On Anthropic's published benchmarks it scores higher on most tasks, and Anthropic says it can match Opus 5.5's top scores for less cost per task at lower effort. Per token it costs twice as much, and Anthropic still recommends Opus 5.5 as the default for most workloads. Test both on your own questions.
Can I turn off thinking on Claude Fable 5.2?
No. Anthropic lists adaptive thinking as always on for Fable 5.2. You steer how much it thinks with the effort setting, which defaults to high on the Claude API.
Can I use Claude Fable 5.2 in Slack?
Yes, through a Slack agent or app that lets you choose the model. In Runbear, save your own Anthropic API key in the agent's settings, choose Claude Fable 5.2 in Model Selection, and connect the agent to a Slack channel.
What is the difference between Claude Fable 5.2 and Claude Mythos 5.2?
Anthropic says they are the same underlying model. Fable adds safeguards for cyber and biology topics and is generally available, while Mythos 5.2 is available only through Anthropic's verification programs and, for enterprise customers, Claude Security.
Related reading
- Claude Enterprise pricing: Team, Enterprise, and API costs
- Claude Haiku 5.5 in Slack: pricing and when to use it
- Best AI agents for Slack in 2026
- Runbear vs Claude: Claude Tag and Claude in Slack compared
Runbear runs shared AI agents in Slack and Microsoft Teams that your team can build without code. See how a no-code Slack AI agent works, or how teams share one agent across a channel.
