Claude Fable 5.2
Model ranking
- Arena Score
- 1478.3425849167
- Arena reference rank
- 26
- Evaluation configuration
- Claude Fable 5.2 (Max)
- Source published
- 2026-11-19
- Votes
- 5216
- Confidence lower
- 1469.6233869722
- Confidence upper
- 1487.0617828612
- Last successful check
- 2026-11-22T14:00:05.877Z
- Last value/rank change
- 2026-11-22T09:30:20.053Z
- Methodology
- Human preference — Text / overall; checked every six hours.
Source: Arena / LMArena ↗ · CC BY 4.0. Selection and catalog matching by XPU LIVE.
| Provider | Price | Provider model ID | Status | Context / max output | Source |
|---|---|---|---|---|---|
Batch / Router catalog Input $4USD / 1M tokensOutput $20USD / 1M tokens Updated 2026-11-22 · 15:43 UTCOpenRouter · Batch / Router catalogInput: $4 Output: $20 Cache Read: $0.1 Cache Write: $5 Web Search: $0.01 Batch processing tariff; not a synchronous Standard request. OpenRouter route catalog price, not a direct-provider price. Routing, modality and endpoint conditions can change the actual bill. Checked 2026-11-22T15:43:15.988Z Price source ↗Router catalog Tariff not listedInput $8USD / 1M tokensOutput $40USD / 1M tokens Updated 2026-11-22 · 15:43 UTCOpenRouter · Router catalogInput: $8 Output: $40 Cache Read: $0.2 Cache Write: $10 Web Search: $0.01 OpenRouter route catalog price, not a direct-provider price. Routing, modality and endpoint conditions can change the actual bill. Checked 2026-11-22T15:43:15.988Z Price source ↗ | anthropic/claude-fable-5.2 | Catalog-listedConditionsPublished standard token tariff. Batch pricing is a separate tariff. Adaptive thinking always on; account eligibility, regional limits and Anthropic retention requirements apply. | Context tokens 1000000Output tokens 128000 | Official source ↗2026-11-10 | |
Standard Tariff not listedInput $8USD / 1M tokensOutput $40USD / 1M tokens Updated 2026-11-22 · 15:43 UTCAnthropic · StandardInput: $8 Output: $40 Cache Write 5m: $10 Cache Write 1h: $16 Cache Read: $0.2 Base USD/MTok rates. Batch, tools and regional premiums may differ. Checked 2026-11-22T15:43:15.988Z Price source ↗ | claude-fable-5-2 | Catalog-listedConditionsPublished standard token tariff. Batch pricing is a separate tariff. Adaptive thinking always on; account eligibility, regional limits and Anthropic retention requirements apply. | Context tokens 1000000Output tokens 128000 | Official source ↗2026-11-10 | |
Standard Tariff not listedInput $8USD / 1M tokensOutput $40USD / 1M tokens Updated 2026-11-22 · 15:43 UTCWaveSpeedAI · StandardInput: $8 Output: $40 Published WaveSpeed token tariff per million tokens. Exact endpoint eligibility and additional charges follow the source. Checked 2026-11-22T15:43:15.988Z Price source ↗ | Not yet captured | Catalog-listedConditionsSource-listed tariff. The live engine supplies prices and observation timestamps. Compare explicit Batch and prompt-length tiers; missing prices are unknown. | Not supplied | Official source ↗2026-11-19 |
Benchmark review · November 19, 2026. The latest verified Artificial Analysis Intelligence Index score is 65 for Claude Fable 5.2 Max, Default Fallback, checked on November 19, 2026. Rank is among verified models in this curated catalog. Benchmark source.
Claude Fable 5.2 is Anthropic’s model for demanding reasoning and long-running agent work. Released on November 9, 2026, it supports text and image inputs and produces text. This page compares its API prices, documented limits and independent benchmark result. You can review direct Anthropic access alongside OpenRouter routes in the provider table.
Claude Fable 5.2 specifications
The official reference lists a one-million-token context window and a maximum output of 128,000 tokens. These limits support long documents, large code reviews and multi-step analysis. Actual request limits can depend on the provider and deployment. Check both the model reference and the endpoint you plan to call.
The direct Claude API identifier is claude-fable-5-2. The OpenRouter identifier is anthropic/claude-fable-5.2. Keeping these identifiers separate helps you select the intended model when moving between services. The model uses proprietary weights, so an API subscription does not provide a downloadable checkpoint.
Reasoning effort and agent workflows
Adaptive thinking is always enabled, and the default effort is high. You can adjust effort to balance reasoning depth, response time and cost. Anthropic positions this release for sustained coding sessions, deep research and work that ends in a finished document or other deliverable.
For an agent workflow, define the required output before the first request. Give the model clear tools and a way to verify completed actions. Then inspect its results on representative tasks. A benchmark score can help shortlist candidates, but your own tests should measure accuracy, task completion and supervision time.
Claude Fable 5.2 API pricing
Anthropic’s standard rate is $8 per million input tokens and $40 per million output tokens. Prompt caching has separate charges. A five-minute cache write costs $10 per million tokens, a one-hour write costs $16, and a cache read costs $0.20. The official Batch API offers a 50% discount on input and output.
The provider table shows the latest collected tariffs and their observation times. OpenRouter’s standard and batch routes have separate identifiers and conditions. Compare the input and output rates from the same tariff. Tool usage, regional processing and other service features can add charges, so review the linked source before estimating a full bill.
Benchmark score and ranking
Claude Fable 5.2 scores 65 on the Artificial Analysis Intelligence Index in the max-effort configuration checked on November 10, 2026. It leads the published model leaderboard in that snapshot. The xpu live comparison table uses one row per base model and selects the strongest verified configuration. Scores tied on the same value follow the site’s alphabetical tie rule.
The xhigh, high, medium and low settings have separate evaluation results. They remain settings of the same model. The price of an API token and the measured benchmark cost per task describe different quantities. Review the independent benchmark profile for the evaluation conditions and the current result.
Access and integration requirements
Anthropic documents a 30-day data-retention requirement for this model unless it expressly authorizes an exception. Forced tool choice is unsupported. Requests should also omit temperature, top_p and top_k overrides. Review these differences when migrating an existing Claude integration.
Claude Mythos 5.2 is a separately named offering for verified organizations. Its access requirements differ. This catalog entry covers Fable and its verified provider routes. See the official specifications and official price list for source details.
Compare alternatives
Compare Claude Fable 5.2 with Claude Opus 5.5 and Claude Sonnet 5.5 using the same prompts and acceptance criteria. Start with a small workload, record the selected reasoning effort, and inspect failures as well as successful answers. Browse the AI model rankings for other candidates and review the comparison methodology for how catalog facts and rankings are presented.
Specifications
- Catalog model ID
- anthropic/claude-fable-5.2
- Developer model ID / revision
- claude-fable-5-2Anthropic model documentation ↗
- Input modalities
- text, imageAnthropic model documentation ↗
- Output modalities
- textAnthropic model documentation ↗
- Context / input token limit
- 1,000,000Anthropic model documentation ↗
- Maximum output tokens
- 128,000Anthropic model documentation ↗
- Tool calling
- Supported
- Structured output
- Not verified in cited sources
- API parameters (source-specific)
- Not verified in cited sources
- Weight access
- Proprietary / hosted APIAnthropic model documentation ↗
- Weight license / access terms
- Anthropic commercial service terms; proprietary weights
- Total parameters (billions)
- Not publicly disclosed
- Active parameters (billions)
- Not publicly disclosed
- Download size (GB)
- Not applicable — no public weights
- Release date
- 2026-11-09Anthropic model documentation ↗
- Notes
- Adaptive thinking is always on; default effort high. Knowledge cutoff June 2026. Forced tool choice and temperature/top_p/top_k overrides are unsupported. Requires 30-day data retention unless expressly authorized. Regional availability and account eligibility apply.
Catalog entries link to their specification and availability sources. This is a curated selection, not every model each provider offers. Best prices are the lowest collected USD tariffs, not a total-cost estimate. Token offers are selected by input price, with output from the same tariff; workload mix may change the cheapest provider. Listed variants may require batch processing, off-peak hours, or specific media settings. Price coverage varies by provider; collected tariffs include their source and observation time. Arena user-preference scores are checked automatically every six hours against the latest published Text / overall dataset. Source: Arena / LMArena; CC BY 4.0. Selection and catalog matching by XPU LIVE. They apply to the shown evaluation configuration; listed prices may use different settings. Models without verified scores are not assumed to be weaker.