Muse Wonder vs Muse Wonder Full
完整资料:Muse Wonder(Meta)和 Muse Wonder Full(Meta),包括全部基准测试、价格和排名。
更新于 · 数据来自 10 个来源
截至 2026年11月16日,Muse Wonder 与 Muse Wonder Full 的质量不相上下(BenchLeader Index 64.4 比 61.7,在 ±3.8 的误差范围内)。在两者都有成绩的所有类别中,Muse Wonder 均领先。
要点(截至 2026年11月16日)
- Muse Wonder 的 BenchLeader Index 为 64.4,Muse Wonder Full 为 61.7(BenchLeader,2026年11月16日)。
- Muse Wonder 每百万 token 混合价格为 $8.00(标价,2026年11月16日)。
按用途怎么选
推荐模型的数值列在前面。类别得分采用 BenchLeader 标尺(50 为模型平均水平,15 分为一个标准差)。领先半个标准差及以上为“高”置信度;相差不到 1 分则视为“均可”。
主要差异
- Muse Wonder 与 Muse Wonder Full (max) 的质量不相上下(64.4 比 61.7,在 ±3.8 的误差范围内)。
- Muse Wonder 在智能体与工具、知识方面更强。
概览
| 规格 | Muse Wonder | Muse Wonder Full |
|---|---|---|
| 提供商 | Meta | Meta |
| 发布日期 | 2026年10月12日 | 2026年10月12日 |
| 上下文窗口(token) | 1,048,576 | 1,000,000 |
| 对比时的配置 | Muse Wonder | Muse Wonder Full (max) |
| 开放权重 | 否 | 否 |
| 每百万 token 价格(输入 / 输出 / 缓存输入) | $4.00 / $20.00 / $0.200 | – |
逐类别对比
类别得分采用 BenchLeader 标尺:50 为参评模型的平均水平,15 分为一个标准差。差距不到 1 分视为持平。
| 类别 | Muse Wonder | Muse Wonder Full (max) | 领先者 |
|---|---|---|---|
| 推理 | – | 65.8 | 无法比较 |
| 编程 | – | 69.9 | 无法比较 |
| 智能体与工具 | 75.6 | 63.6 | Muse Wonder |
| 数学 | 70.1 | – | 无法比较 |
| 知识 | 77.1 | 62.3 | Muse Wonder |
| 人类偏好 | 69.6 | – | 无法比较 |
| 长上下文 | – | 66.6 | 无法比较 |
同一图表对比
各领域采用与 Index 相同的标尺,平均模型为 50 分。单击某个领域可查看构成它的基准测试。 可从列表中最多选择 5 个模型。
虚线:该模型在所经过的轴上没有结果,线条只连接相邻的点。
未绘制(这些模型没有结果):指令遵循, 多模态。
各来源的分歧
在两者都有结果的 0 项基准测试中,Muse Wonder 领先 0 项,Muse Wonder Full 领先 0 项。所有同时评测两者的综合评分都认为 Muse Wonder Full 更强。
| 综合 Index | Muse Wonder | Muse Wonder Full (max) | 领先者 |
|---|---|---|---|
| BenchLeader Index | 64.4 | 61.7 | 不相上下 |
| AA Intelligence Index v5.0.1 | 40.5† | 50.1 | Muse Wonder Full |
| Vals Index | 64.9† | 66.8 | Muse Wonder Full |
各发布方使用各自的任务、评测框架和推理强度设置,因此总体结论可能不同;BenchLeader 将它们统一到同一标尺上。† 表示同一模型在另一推理强度设置下的测量值(发布方未测试本页对比的设置)。
价格、速度与延迟
标价按每 1 个输出 token 对应 3 个输入 token 混合计算。每项任务的价格采用方法论页面中的 token 数量。每项基准测试任务的成本、token 数和耗时均为实测值,取自 Artificial Analysis 整个 Intelligence Index 测试的平均值。
| 指标 | Muse Wonder | Muse Wonder Full (max) |
|---|---|---|
| 每百万 token 混合价格 | $8.00 | – |
| 每百万 token 输入 / 输出 | $4.00 / $20.00 | – / – |
| 每百万 token 缓存输入 | $0.200 | – |
| 每项任务价格:聊天回复 | $0.0076 | – |
| 每项任务价格:总结一份 30 页的报告 | $0.060 | – |
| 每项任务价格:代码编辑 | $0.054 | – |
| 每项任务价格:智能体编程会话 | $0.320 | – |
| 每项任务价格:结构化信息提取 | $0.012 | – |
| 智能体编程会话(输入大多命中缓存) | $0.149 | – |
| 每项任务输出 token(AA Index) | – | 88k |
| 其中推理部分 | – | 55k |
| 输出速度 | 70 tok/s | – |
| 首次回答时间 | 3.3 秒 | – |
每个模型都按其排名最高的推理强度设置进行对比。Muse Wonder 得分 64.4,首个回答 token 需 3 秒;Muse Wonder (max) 为 60.8(20 秒),Muse Wonder (high) 为 56.2(13 秒)。Muse Wonder Full (max) 得分 61.7。推理强度越高,回答前思考越久:得分更高,但等待更长、计费 token 也更多。 推理强度如何影响得分和成本
全部基准测试
共 130 项基准测试,按类别分组;点击名称可查看该排行榜及领先模型。† 表示同一模型在另一推理强度设置下的测量值,仅供参考,不计入胜场。
| 指标 | 发布方 | Muse Wonder | Muse Wonder Full (max) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 64.4 | 61.7 |
| 智能体与工具得分 | BenchLeader | 75.6 | 63.6 |
| 人类偏好得分 | BenchLeader | 69.6 | – |
| 知识得分 | BenchLeader | 77.1 | 62.3 |
| 数学得分 | BenchLeader | 70.1 | – |
| 编程得分 | BenchLeader | – | 69.9 |
| 长上下文得分 | BenchLeader | – | 66.6 |
| 推理得分 | BenchLeader | – | 65.8 |
| 混合价格(美元/百万 token) | BenchLeader | $8.00 | – |
| 输出速度 | BenchLeader | 70 tok/s | – |
| 首次回答时间 | BenchLeader | 3.3 秒 | – |
| 上下文窗口 | BenchLeader | 1,048,576 token | 1,000,000 token |
| 推理 | |||
| Arena Hard Prompts | Arena (LMArena) | 1531† | – |
| LiveBench 2.0 Reasoning | LiveBench | 90.6%† | – |
| LiveBench 1.0 Reasoning | LiveBench | 90.6%† | – |
| Humanity's Last Exam (AA) | Artificial Analysis | 50.1%† | 53.6% |
| HLE-Diamond (AA) | Artificial Analysis | 39.0%† | 48.5% |
| CritPt | Artificial Analysis | 28.3%† | 32.9% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 50.5%† | 56.5% |
| MysteryMechanism | Vals AI | 27.5%† | 35.1% |
| HLE-Diamond (Scale) | Scale AI SEAL | 37.3% | – |
| Arena Expert | Arena (LMArena) | 1525† | – |
| Mystery Game Puzzles | Epoch AI Benchmarking Hub | 51.0%† | – |
| LMCA | Epoch AI Benchmarking Hub | 56.8% | – |
| DTBench | Epoch AI Benchmarking Hub | 95.2% | – |
| 编程 | |||
| SciCode | SciCode | 60.3%† | – |
| Arena Coding | Arena (LMArena) | 1542† | – |
| Arena WebDev | Arena (LMArena) | 1647† | 1724 |
| LiveBench 2.0 Coding | LiveBench | 74.9%† | – |
| LiveBench 1.0 Coding | LiveBench | 80.7%† | – |
| SciCode (AA) | Artificial Analysis | 60.3%† | 62.6% |
| IOI | Vals AI | 66.9%† | 79.8% |
| Code Migration | Vals AI | 62.5%† | 61.0% |
| Code Migration v1.1 | Vals AI | 72.4%† | 62.8% |
| ProgramBench | Vals AI | – | 0.5% |
| SWE-bench Multimodal | Vals AI | 45.6%† | 47.3% |
| Vibe Code Bench 1-100 | Vals AI | 26.6%† | 34.0% |
显示全部 130 项基准测试
| 指标 | 发布方 | Muse Wonder | Muse Wonder Full (max) |
|---|---|---|---|
| 编程 | |||
| Vibe Code Bench v1.1 | Vals AI | 83.1%† | 86.0% |
| SWE Atlas v1.1 | Scale AI SEAL | 64.8% | – |
| SWE Atlas v1.1: Codebase QnA | Scale AI SEAL | 73.4% | – |
| SWE Atlas v1.1: Test Writing | Scale AI SEAL | 61.1% | – |
| SWE Atlas v1.1: Refactoring | Scale AI SEAL | 60.0% | – |
| SWE-Bench Pro V2 | Scale AI SEAL | 96.7%† | – |
| SWE-Bench Pro V2 HARD | Scale AI SEAL | 90.2%† | – |
| MirrorCode | Epoch AI Benchmarking Hub | 45.6%† | – |
| CursorBench | Cursor | 50.4%† | – |
| WeirdML v3 | WeirdML | 15.2%† | – |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 52.0%† | 57.1% |
| GameBench 2 Overall | DuelLab GameBench 2 | 64.8 | – |
| GameBench 2 by reasoning setting | DuelLab GameBench 2 | 54.9† | – |
| Design Arena Website | Design Arena | 1378† | – |
| FrontierCode 1.1 (AA) | Artificial Analysis | 46.5%† | 44.9% |
| FrontierCode 2.0 (AA) | Artificial Analysis | 27.9%† | 37.8% |
| 智能体与工具 | |||
| Terminal-Bench | Terminal-Bench | 47.0%† | – |
| APEX-Agents | Mercor | 66.1%† | – |
| Arena Agent | Arena (LMArena) | 1.8† | 9 |
| LiveBench Long Agentic | LiveBench | 46.2%† | – |
| LiveBench Browser | LiveBench | 60.0%† | – |
| LiveBench Agentic Coding | LiveBench | 62.7%† | – |
| HiL-Bench | Scale AI SEAL | 79.7% | – |
| GDPval-AA v2.1, normalised score | Artificial Analysis | 63.5%† | 67.8% |
| GDPval-AA v2.1 | Artificial Analysis | 1770† | 1857 |
| CUA-bench | Vals AI | 7.2%† | – |
| CyberBench | Vals AI | 80.4%† | 73.6% |
| Finance Agent v2 | Vals AI | 56.6%† | 59.1% |
| Finance Agent v2.1 | Vals AI | 65.5%† | 67.6% |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 23.8%† | 23.3% |
| Harvey's Legal Agent Benchmark v1 | Vals AI | 17.1%† | 21.7% |
| Legal Research Bench | Vals AI | 50.5%† | 53.4% |
| Robo-bench | Vals AI | – | 10.4% |
| SREBench | Vals AI | – | 11.8% |
| Tax Agent Bench | Vals AI | 36.3%† | 43.0% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 49.5%† | 58.1% |
| Terminal-Bench Science | Vals AI | 22.9%† | 37.1% |
| Arena Agent Task Completion | Arena (LMArena) | 5.4† | 11.1 |
| PostTrainBench | Epoch AI Benchmarking Hub | 24.2%† | – |
| Vending-Bench 2 | Andon Labs | 9432.2 | – |
| AutomationBench | Artificial Analysis | 68.1%† | 71.3% |
| GDP.pdf (AA) | Artificial Analysis | 30.6%† | 32.6% |
| Harvey LAB-AA v1.1 | Harvey | 10.0%† | 13.1% |
| AA-Briefcase v1.1 | Artificial Analysis | 1659† | 1729 |
| GDPx | GDPx by SemiAnalysis | 6.8%† | 9.0% |
| Design Arena Agentic | Design Arena | 1177† | – |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 25.7%† | 37.1% |
| MKA (AA) | Artificial Analysis | 47.5%† | – |
| 数学 | |||
| FrontierMath Tiers 1–3 | Epoch AI Benchmarking Hub | 82.1%† | – |
| FrontierMath Tier 4 | Epoch AI Benchmarking Hub | 70.7%† | – |
| OTIS Mock AIME | Epoch AI Benchmarking Hub | 99.4%† | – |
| ProofBench | Vals AI | 84.0% | – |
| LiveBench Mathematics | LiveBench | 96.5%† | – |
| Arena Maths | Arena (LMArena) | 1490† | – |
| 知识 | |||
| LiveBench 2.0 Data Analysis | LiveBench | 81.6%† | – |
| LiveBench Research | LiveBench | 93.0%† | – |
| LiveBench 1.0 Data Analysis | LiveBench | 81.6%† | – |
| AA-Omniscience v1.1 | Artificial Analysis | 25.1† | 30.6 |
| LegalBench | Vals AI | – | 87.4% |
| PRBench Finance | Scale AI SEAL | 58.7% | – |
| PRBench Legal | Scale AI SEAL | 58.4% | – |
| BioMysteryBench | Vals AI | 67.0%† | 73.3% |
| Excel Modeling Benchmark | Vals AI | 67.1%† | 71.9% |
| Excel Modeling Benchmark v1.1 | Vals AI | 72.5%† | 75.0% |
| MedCode | Vals AI | – | 49.9% |
| MedScribe | Vals AI | – | 89.3% |
| Public Benefits Bench v1.1 | Vals AI | 70.2%† | 73.8% |
| Access to Justice Benchmark | Vals AI | 73.0%† | 74.7% |
| Financial Coaching Bench | Vals AI | 73.8%† | 75.0% |
| Immigration Bench | Vals AI | 69.6%† | 71.2% |
| EBR-bench | Epoch AI Benchmarking Hub | 31.4%† | – |
| GDP.pdf | Epoch AI Benchmarking Hub | 30.8%† | – |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 38.0%† | 41.0% |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 79.1%† | 82.4% |
| 指令遵循 | |||
| LiveBench Language | LiveBench | 82.9%† | – |
| LiveBench Instruction Following | LiveBench | 73.0%† | – |
| Arena Instruction Following | Arena (LMArena) | 1498† | – |
| 人类偏好 | |||
| Arena Text | Arena (LMArena) | 1511† | – |
| EQ-Bench 4 | EQ-Bench | 1325 | – |
| EQ-Bench Creative Writing v3 | EQ-Bench | 2007 | – |
| Arena Creative Writing | Arena (LMArena) | 1491† | – |
| Arena Multi-turn | Arena (LMArena) | 1525† | – |
| Arena Longer Queries | Arena (LMArena) | 1507† | – |
| 多模态 | |||
| LiveBench Design | LiveBench | 70.9%† | – |
| MMMU-Pro | Artificial Analysis | 83.8%† | – |
| Humanity's Sixth Sense | Scale AI SEAL | 43.7%† | – |
| Blueprint-Bench 2 | Epoch AI Benchmarking Hub | 31.7% | – |
| Furniture Assembly | Epoch AI Benchmarking Hub | 54.2%† | – |
| VIALS-AA | Handshake AI | 55.3%† | – |
| Code2Video-AA | Artificial Analysis | 997† | – |
| 长上下文 | |||
| AA-LCR v1.1 | Artificial Analysis | 80.7%† | 83.7% |
| MLCR | Artificial Analysis | 22.2%† | 22.2% |
| 综合 | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | 161.7 | – |
| LiveBench 2.0 | LiveBench | 70.7%† | – |
| LiveBench 1.0 | LiveBench | 81.2%† | – |
| AA Intelligence Index v5.0.1 | Artificial Analysis | 43.6† | 50.1 |
| AA Intelligence Index v4.3.2 | Artificial Analysis | 52.4† | 55.8 |
| AA Progress Index v1 | Artificial Analysis | 335.8† | 343.2 |
| Vals Index | Vals AI | 64.9† | 66.8 |
| Vals Finance Index | Vals AI | 61.2%† | 64.5% |
| Vals RSI Index | Vals AI | – | 24.4% |
| Vals Science Index | Vals AI | 39.1%† | 48.5% |
| Vals Router Index | Vals AI | 64.9% | 66.8%† |
| Vals Legal Index | Vals AI | 75.3%† | 77.9% |
| 安全与诚实 | |||
| DistressBench | Scale AI SEAL | 85.9% | – |
数据截至 2026年11月16日。对比的是每个模型的最佳配置;每项分数的来源链接都在模型页面上。
DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.
变化
- Muse Wonder:BenchLeader Index 63.0 → 64.4。
- Muse Wonder:BenchLeader Index 63.5 → 63.0。
- Muse Wonder Full (max):BenchLeader Index 62.6 → 61.7。
- Muse Wonder:1 项新结果(Blueprint-Bench 2)。
- Muse Wonder:BenchLeader Index 64.3 → 63.5。
- Muse Wonder:5 项分数修订。
- Muse Wonder:1 项分数修订。
- Muse Wonder Full (max):1 项分数修订。
- Muse Wonder Full (max):1 项分数修订。
- Muse Wonder Full (max):1 项新结果(Arena Agent)。
相关对比
- Muse Wonder vs Claude Fable 5.2
- Muse Wonder vs Gemini 4 Carbon
- Muse Wonder vs Claude Opus 5.5
- Muse Wonder Full vs Claude Fable 5.2
- Muse Wonder Full vs Gemini 4 Carbon
- Muse Wonder Full vs Claude Opus 5.5
- Muse Wonder Full vs GPT-6 Astra
- Muse Spark 1.3 vs Muse Wonder Full
排行榜: 综合最佳大模型 · 最适合编程的大模型 · 最适合智能体的大模型 · 便宜好用的大模型
Muse Wonder vs Muse Wonder Full:常见问题
- Muse Wonder 比 Muse Wonder Full 更好吗?
- Muse Wonder 与 Muse Wonder Full (max) 的质量不相上下(64.4 比 61.7,在 ±3.8 的误差范围内)。BenchLeader Index 综合了所有第三方质量基准测试;截至 2026年11月16日,两者差距在误差范围内,请结合你的用途查看各类别得分。
- 智能体和工具调用用 Muse Wonder 还是 Muse Wonder Full 更好?
- Muse Wonder 在智能体任务方面得分更高(类别得分 76 比 64,50 为平均水平)。
- 处理长文档哪个更好?
- 两者都约为 1M token(Muse Wonder 为 1,048,576,Muse Wonder Full 为 1,000,000)。
- Muse Wonder 是什么时候发布的?
- Muse Wonder 于 2026年10月12日 发布。