Anthropic has released Claude Fable 5, its latest-generation AI model — and the first to break away from the Claude 4 naming line. Fable 5 (model ID claude-fable-5) arrives as the newest and most capable model in Anthropic’s catalog, sitting alongside the Claude 4.x family — Opus 4.8, Sonnet 4.6 and Haiku 4.5 — which remains available. Here is what we know, what the new name means, and what it changes for everyday Claude users.
What is Claude Fable 5?
Fable 5 is Anthropic’s new model generation, available through the Claude API and in products like Claude Code (Anthropic’s coding agent for terminal, desktop and web). Its developer identifier is claude-fable-5, and Anthropic positions it as the go-to model for the hardest tasks: complex reasoning, agentic coding and long-context work.
The most striking detail is the name itself. Until now, Anthropic organized its models into three fixed sizes — Haiku, Sonnet and Opus, smallest to largest. “Fable” introduces a brand-new name, suggesting a generation that stands apart from the 4.x line rather than a simple iteration.
Anthropic’s model lineup now
| Model | Model ID | Position |
|---|---|---|
| Claude Fable 5 | claude-fable-5 | Latest generation, newest release |
| Claude Opus 4.8 | claude-opus-4-8 | Top of the 4.x family |
| Claude Sonnet 4.6 | claude-sonnet-4-6 | Capability/cost balance |
| Claude Haiku 4.5 | claude-haiku-4-5 | Fast and affordable |
Anthropic’s guidance for builders is to default to its latest and most capable models. The 4.x family is not going away: it remains the choice when cost matters most (Haiku) or when you already have validated workflows on Sonnet or Opus. On the speed front, Claude Code also offers a “fast mode” on Opus that accelerates output without dropping to a smaller model.
What it means for you
If you use Claude as an assistant
Generational jumps show up most in the hard stuff: long instructions with many conditions, lengthy documents, multi-step tasks. If Claude is already your assistant, Fable 5 is the version to try first on exactly those tasks.
If you build with the API
The usual playbook applies: run your evals against claude-fable-5 before migrating production, and keep 4.x as a fallback while you compare quality and cost on your real workload. If you build agents — the area Anthropic has pushed hardest with Claude Code and the Agent SDK — that is where the upgrade makes the most sense.
If you compete in SEO/GEO
Every model jump changes how AI assistants read and cite the web. Content with direct answers, concrete data and clean structure remains what these models prefer to cite — the strategy in our GEO guide doesn’t change; it gets reinforced.
What we don’t know yet
As of today, Anthropic has not publicly detailed every Fable 5 specification (family sizes, per-token pricing, full benchmark numbers). We will update this piece as official data lands. In the meantime, be skeptical of specific figures circulating without a source — that is the classic pattern in the days after a launch.
Claude Fable 5 FAQ
Can I use Claude Fable 5 today?
Yes: it is available via the Anthropic API under the ID claude-fable-5 and in Anthropic products like Claude Code. In the Claude app, availability depends on your plan.
Does it replace Opus 4.8?
Not immediately — they coexist. Fable 5 is the new generation; Opus 4.8 remains the top of the 4.x family and a valid production choice.
Why the name “Fable”?
Anthropic has not officially explained the choice. What matters is that it breaks the Haiku/Sonnet/Opus scheme — a signal of a new generation with its own identity.
Is it better than GPT or Gemini?
Without full public benchmarks, any categorical claim is premature. The honest answer: test it on your real tasks and compare — which is what we do in our comparisons, to be updated with Fable 5.
Does anything change for free users?
Historically, each new generation eventually improves what free plans get, with limits. If you use Claude for free, watch the coming weeks.
Bottom line
- Fable 5 is Anthropic’s new most capable model, with the ID
claude-fable-5 - It coexists with the Claude 4.x family (Opus 4.8, Sonnet 4.6, Haiku 4.5)
- Available via API and Claude Code; try it first on your hardest tasks
- Full specifications not yet published — we will update with official data
More context in how to use Claude and our ChatGPT vs Claude comparison.
