Fable 5.2 mode
What Claude Fable 5.2 can do
Anthropic released Claude Fable 5.2 on November 9, 2026. We run it through the same published briefs every model in the arena gets, one shot per effort level, and post each output as soon as it finishes. Every card opens the real page exactly as the model wrote it, with nothing fixed afterwards.
- Price per 1M tokens
- $8 in / $40 out
- Fable 5.1 was $10 / $50
- Effort levels
- low to max
- Claude apps default to medium
- Outputs live
- 41
- Across 21 challenges
One landing page, five effort levels
Same brief (Driftwood, a travel photo book SaaS page), same harness; only the effort flag changes. Cost is priced at API list rates and tokens include thinking.
Same brief: Fable 5.2 vs Fable 5.1
| Effort | Tokens 5.2 | Tokens 5.1 | Change | Cost 5.2 | Cost 5.1 |
|---|---|---|---|---|---|
| low | 33.6k | 27.4k | +23% | $1.35 | $2.17 |
| medium | 51.5k | 26.5k | +94% | $2.08 | $2.10 |
| high | 117.7k | 29.7k | +297% | $4.72 | $2.25 |
| xhigh | 175.2k | 95.7k | +83% | $9.07 | $5.58 |
| max | 226.5k | 144.3k | +57% | $9.09 | $10.12 |
Tokens are output tokens, thinking included. The Fable 5.1 costs reported by the CLI carry about $1.21 of input-side cost per run, against $0.42 for these Fable 5.2 runs, so compare tokens first.
Every Fable 5.2 output so far
Filter by effort level or challenge. New runs show up here as they finish.
1 · Landing
2 · Game
3 · Tool
4 · Dashboard
5 · Particles
6 · 3D Game
7 · CHIP-8
8 · News
12 · Digger
13 · Artillery
14 · Tower Def
22 · Fluid
26 · Traffic
30 · Charts
31 · Music
32 · Spreadsheet
33 · Kanban
34 · Editor
35 · Diagram
36 · Pixel Art
43 · Chess
How these were made
Each challenge and effort level runs once through the Claude Code CLI with --effort set per run, no tools and no cherry-picking. The briefs are the same ones every model in the arena got.
Cost is the run priced at API list rates ($8 / $40 per 1M tokens), as reported by the CLI. Time is wall-clock generation time.
/method