Claude Fable 5.2 Is Here: What It Changes for the Forms Your Agent Builds (2026)
Claude Fable 5.2 is Anthropic's new top model, now cheaper than Fable 5.1. What it changes when your agent builds, scores and reads forms.
Josue HernandezAI Specialist
Claude Fable 5.2 is Anthropic's newest and most capable model. On November 9, 2026, Anthropic launched Claude Fable 5.2 and Claude Mythos 5.2, two versions of the same model. Fable is for everyone and keeps strong safeguards. Mythos drops some of them and only goes to verified organizations.
If your agent builds forms, scores leads or reads responses, the model behind it matters. So here is what changed, which numbers come from Anthropic, and when the new model is worth it for form work.
What Anthropic launched
Here are the facts from the launch post and the Claude Platform release notes, and nothing more:
- Price: $8 per million input tokens and $40 per million output tokens. Fable 5.1 cost $10 and $50, so the new model is cheaper.
- Fast mode: up to 2.5 times faster output, at $16 and $80 per million tokens.
- Context: a 1M token context window and up to 128k output tokens, with adaptive thinking always on.
- Speed: Anthropic says it writes output up to 30% faster than Fable 5.1.
- Where: the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry. Pro plans and standard Team seats can use Fable for up to 50% of their weekly usage limits.
- Model ID:
claude-fable-5-2.
The numbers that matter for forms
Most of the launch is about coding, science and security. However, two rows in Anthropic's benchmark table are close to form work:
| Benchmark (Anthropic's table) | Fable 5.2 | Fable 5.1 | Opus 5.5 |
|---|---|---|---|
| AutomationBench (business workflows) | 51.4% | 31.4% | 42.5% |
| OSWorld 2.1 (computer use, offline set, partial) | 89.3% | 83.5% | 87.2% |
Business workflows are where forms live: intake, qualification, routing to a CRM. In that row, the new model jumps 20 points over Fable 5.1. Still, Anthropic's own table puts Gemini 4 Carbon at 53.9% on the same test, so the lead is not universal. These are Anthropic's numbers, not independent tests.
What Claude Fable 5.2 changes for your agent's forms
Bigger jobs in one run. A 1M token window means your agent can read a large batch of form responses at once. For example, it can review a month of answers, compare scores and tell you which questions people skip.
A planner on top of cheaper models. Anthropic says Fable 5.2 can act as an Advisor, consulted by Haiku, Sonnet or Opus models, or as an Orchestrator that hands simpler tasks to smaller models. In practice, the expensive model can design the scoring rules while a cheaper one builds the form.
But you probably do not need it for every form. Anthropic itself recommends that most users "start with Claude Opus 5.5, our strongest daily driver." A lead form with ten questions is not open-ended research. So keep Fable for the hard parts: a scoring model with many rules, a multi-step screening flow, or an analysis across hundreds of responses.
Try it with dForms
dForms is the form builder for AI agents. Your agent builds the form through the API or the dForms MCP at mcp.dforms.ai, publishes it, and scores every answer. Then dForms can send the lead to your calendar, HubSpot or a webhook.
The model you pick in Claude is the one that calls the connector. So pick Fable 5.2 for the scoring design, then paste this prompt:
Design and build a lead qualification form in dForms for our B2B agency.
Ask for name, work email, company, team size, monthly budget,
timeline to start, and the main problem they want to solve.
Write a scoring rule for every answer, with more points for a budget
over $2,000 a month, a start date within 30 days and a team of 10 or more.
Explain each rule in one line so I can check it.
Leave the form in draft so I can review it before you publish.
Your agent builds the form as a draft. You review the rules, then publish. If you change your mind later, a published form can be reverted. For the scoring side, see our guide to a form with scoring. To set up the connector first, read how to create a form with Claude, or start free on dForms.
Frequently asked questions
What is the difference between Claude Fable 5.2 and Claude Mythos 5.2?
They are the same underlying model. According to Anthropic, Fable 5.2 has strong cybersecurity and biology safeguards and is open to all users. Mythos 5.2 has the same capabilities without those safeguards, and only verified organizations can use it.
How much does Fable 5.2 cost on the API?
It costs $8 per million input tokens and $40 per million output tokens on the API. That is less than Fable 5.1, which cost $10 and $50, per the release notes.
Should my agent use Claude Fable 5.2 to build forms?
Only for the hard parts. Anthropic recommends Claude Opus 5.5 for most tasks and Fable for the most ambitious work. A simple form runs fine on a daily model, while complex scoring or large response analysis is where Fable earns its price.
