Anthropic launches Fable 5.2 on Claude Platform and Claude Code

Nov 10

1. 12:45

   heise onlineRelated discussion

   heise online joined this discussion ↗

   This post contributes to heat; it is not a formal report or confirmed fact.

2. 12:21

   English text is being prepared.Related discussion

   English text is being prepared. joined this discussion ↗

   This post contributes to heat; it is not a formal report or confirmed fact.

3. 07:53

   KazikeSelected

   Hands-on with Claude Fable 5.2: Outstanding Aesthetics and Detail

   Anthropic released Claude Fable 5.2 and launched it across all channels, retaining a 1M Token context and 128K maximum output, with pricing reduced from $10 per million input Tokens and $50 per million output Tokens to $8 input and $40 output, an overall price cut of 20%. As a $200 Ultra member, the author tested about 10 tasks, using one-shot prompts to generate a ChatGPT fourth-anniversary tribute video, a Three.js 3D scene, and a 3D web page for Apple Jing'an, concluding that it is clearly stronger than Opus 5.5 in aesthetics and detail, but that it consumed about 24% of the quota in nearly 4 hours, making the cost relatively high.

   Expand summary
4. 07:42

   David OndrejRelated discussion

   David Ondrej joined this discussion ↗

   This post contributes to heat; it is not a formal report or confirmed fact.

5. 07:21

   Oran GeRelated discussion

   Oran Ge joined this discussion ↗

   This post contributes to heat; it is not a formal report or confirmed fact.

6. 06:08

   QbitAISelected

   Anthropic Releases Claude Fable 5.2 and Mythos 5.2, with the High Tier Surpassing in Cost-Effectiveness…

   Anthropic released Claude Fable 5.2 and Mythos 5.2. The company claims that 1,000 Agents running continuously for 8 days autonomously wrote a complete, bootable operating system from an empty folder. Compared with 5.1, the new models generate 30% faster, and API pricing has been cut by 20% to $8 per million input tokens and $40 per million output tokens; a 2.5x-speed Fast mode is also available at double the price. On FrontierCode 2.0, Fable 5.2's High tier scores 45.3% at a cost of $6.72 per task, close to Opus 5.5 Max tier's 45.5% and $12.82; the company recommends using models such as Opus 5.5 for everyday work and handing difficult problems to Fable 5.2 as an advisor or orchestrator. The technical documentation also discloses that during training, two capabilities—accelerator kernel optimization and model architecture improvements—were locked at the Opus 5.5 level, and that when high-risk cybersecurity or biological interception is triggered, it falls back in the background to the older model to answer.

   Expand summary
7. 05:04

   Daniel MiesslerRelated discussion

   Daniel Miessler joined this discussion ↗

   This post contributes to heat; it is not a formal report or confirmed fact.

8. 04:17

   TeortaxesRelated discussion

   Teortaxes joined this discussion ↗

   This post contributes to heat; it is not a formal report or confirmed fact.

9. 04:04

   Hacker News Trending

   Artificial Analysis evaluates Claude Fable 5.2: Intelligence Index 65…

   Artificial Analysis released its evaluation of Claude Fable 5.2 (Max, Default Fallback). The model scores 65 on the Artificial Analysis Intelligence Index, above the median of 40 for models at the same price tier. Its input price is $8 per 1M tokens and output is $40, with an average cost of $26.54 per Intelligence Index task, 45.0 minutes per task, and a first-answer token latency of 171.16s. The model is a reasoning model that supports text and image input and text output, with a 1M-token context window; its weights are not public.

   Expand summary
10. 03:52

    CursorRelated discussion

    Cursor joined this discussion ↗

    This post contributes to heat; it is not a formal report or confirmed fact.

11. 03:45

    Artificial AnalysisSelected

    Anthropic releases Claude Fable 5.2, Artificial Analysis Intelligence Index…

    Anthropic releases Claude Fable 5.2, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 65 (max effort), leading Gemini 4 Carbon (high, 60) by 5 points and Claude Opus 5.5 (max, 59) by 6 points. The improvements mainly come from scientific reasoning, knowledge, and FrontierCode 2.0, reaching 80% on HLE-Diamond (the previous best was 69% for Gemini 4 Carbon). Pricing is $8/$40 per 1M input/output tokens, 20% lower than Fable 5.1, but at max effort the average cost per task is $26.5, about 2.6 times that of Opus 5.5 ($10.0); the context window is 1 million tokens, supports image and text input, and offers five effort settings.

    Expand summary
12. 03:04

    Claude DevsFirst-partySelected

    Claude Platform and Claude apps announce default effort level settings

    Claude Devs explains that effort defaults to high on Claude Platform and medium in Claude apps. xhigh is reserved for the hardest long-running agent tasks and requires measurable gains; Max is used to probe capability limits and is not recommended for everyday use.

    Expand summary
13. 03:04

    Claude DevsFirst-partySelected

    Claude Fable 5.2 launches on Claude Platform and Claude Code

    Claude Fable 5.2 is now available on Claude Platform and Claude Code, up to 30% faster than Fable 5.1, priced at $8/$40 per 1M tokens, down from the previous $10/$50. effort defaults to high on Claude Platform and medium in the Claude app, with xhigh reserved for the hardest long-running agent tasks. Fable 5.2 also offers a Fast mode with 2.5x faster output at double the price; set "speed": "fast" in the API, and in Claude Code you can switch it in the model selector.

    Expand summary