Model dossier模型档案
Claude Fable 5
Claude Fable 5 is a frontier-tier model from Anthropic (Intelligence Score 94.4/100). Strongest in agent (99.4) and math (99.3); best fit: tool-use and multi-step agents. It trails in preference (51.8). Premium-priced: $10 in / $50 out per 1M tokens. Context window 1M tokens — long-document friendly.
Claude Fable 5 是 Anthropic 的前沿梯队模型(智能评分 94.4/100)。 最强项是智能体(99.4分),其次是数学(99.3分);适合工具调用与多步 Agent。 短板是偏好(51.8分)。 定价属高端定价:每百万 token 输入 $10 / 输出 $50。 上下文 1M token,长文档友好。
94.4
Intelligence Score · benchmark coverage 智能评分 · benchmark 覆盖 95%
◈
Benchmark scoresBenchmark 成绩
10 entries 条| Benchmark基准 | Score成绩 | Index指数 | Dated日期 | Measured by测评方 | |
|---|---|---|---|---|---|
| Humanity's Last ExamReasoning | 53.3 | 93 | — | official官方榜 | source ↗×5 |
| ↳ ⚡max | 53.3 | — | — | 3rd-party第三方 | source ↗ |
| GPQA DiamondReasoning | 94.1⚡max | 99 | — | 3rd-party第三方 | source ↗×8⚠±15.7 |
| MMLU-ProKnowledge | 91.5 | 96 | 2026-08-19 | 3rd-party第三方 | source ↗×2 |
| SWE-bench VerifiedCoding | 95 | 99 | 2026-08 | 3rd-party第三方 | source ↗×5 |
| AIME 2025Math | 99.7 | 100 | — | 3rd-party第三方 | source ↗×3⚠±12.7 |
| LiveCodeBenchCoding | 89.78 | 96 | — | 3rd-party第三方 | source ↗×3⚠±9.48 |
| FrontierMathMath | 88 | 99 | — | official官方榜 | source ↗×5⚠±4.6 |
| Terminal-BenchCoding | 83.8 | 91 | — | official官方榜 | source ↗×6⚠±53.9 |
| τ²-benchAgent | 98.5 | 99 | — | 3rd-party第三方 | source ↗×4 |
| LMArena (Chatbot Arena)Preference | 1506 | 52 | 2026-08-12 | official官方榜 | source ↗×9 |
≡
Specs & pricing规格与价格
| Provider厂商 | Anthropic |
|---|---|
| Released发布 | 2026-06-09 |
| Context window上下文 | 1M tokens |
| Max output最大输出 | 128K tokens |
| Modality模态 | text+image+file->text |
| Reasoning tiers推理档位 | ⚡low ⚡medium ⚡high ⚡max default默认 high · low / medium / high / x-high |
| API input price输入价格 | $10 / 1M tokens |
| API output price输出价格 | $50 / 1M tokens |
⛓
Tools using this model使用它的工具
How we score →评分方法 → · Catalog & pricing via the official OpenRouter API.模型目录与价格来自 OpenRouter 官方 API。