人工智能模型能力比較
知識工作
Only models released within the past year are shown. Only benchmarks covering at least 3 of these models are shown. Unknown release dates are excluded; historical data is retained.
— means no published score for this candidate in this source, not zero.
| 編號 | 模型 | 10/11 recent-year models have scores | 10/11 recent-year models have scores | 10/11 recent-year models have scores | 10/11 recent-year models have scores | 10/11 recent-year models have scores | 9/11 recent-year models have scores |
|---|---|---|---|---|---|---|---|
| 1 |  Claude Opus 5.5 | 65.6maxClaude Opus 5.5 (Max, Default Fallback) | 60.4maxClaude Opus 5.5 (Max, Default Fallback) | 60.7maxClaude Opus 5.5 (Max, Default Fallback) | 63.3maxClaude Opus 5.5 (Max, Default Fallback) | 63.7maxClaude Opus 5.5 (Max, Default Fallback) | 60.5maxClaude Opus 5.5 (Max, Default Fallback) |
| 2 |  GPT-6 Astra | 59.8maxGPT-6 Astra (Max) | 54.7maxGPT-6 Astra (Max) | 55.2maxGPT-6 Astra (Max) | 58.5maxGPT-6 Astra (Max) | 57.3maxGPT-6 Astra (Max) | 51.7maxGPT-6 Astra (Max) |
| 3 | GPT-6.1 Sol | 59.3maxGPT-6.1 Sol (Max) | 53.9maxGPT-6.1 Sol (Max) | 54maxGPT-6.1 Sol (Max) | 57.7maxGPT-6.1 Sol (Max) | 57maxGPT-6.1 Sol (Max) | 51.2maxGPT-6.1 Sol (Max) |
| 4 | Claude Fable 5.1 | 62.7maxClaude Fable 5.1 (Max, Default Fallback) | 56.5maxClaude Fable 5.1 (Max, Default Fallback) | 56.4maxClaude Fable 5.1 (Max, Default Fallback) | 60.5maxClaude Fable 5.1 (Max, Default Fallback) | 59.7maxClaude Fable 5.1 (Max, Default Fallback) | 58maxClaude Fable 5.1 (Max, Default Fallback) |
| 5 |  Gemini 4 Argon | 60.4highGemini 4 Argon (High) | 55.2highGemini 4 Argon (High) | 53.8highGemini 4 Argon (High) | 60highGemini 4 Argon (High) | 56.2highGemini 4 Argon (High) | — |
| 6 | Claude Sonnet 5.5 | 60.8maxClaude Sonnet 5.5 (Max, Default Fallback) | 58.4maxClaude Sonnet 5.5 (Max, Default Fallback) | 57.2maxClaude Sonnet 5.5 (Max, Default Fallback) | 56.4maxClaude Sonnet 5.5 (Max, Default Fallback) | 59.6maxClaude Sonnet 5.5 (Max, Default Fallback) | 58.1maxClaude Sonnet 5.5 (Max, Default Fallback) |
| 7 | Claude Opus 5 | 60.8maxClaude Opus 5 (Max) | 54maxClaude Opus 5 (Max) | 54.9maxClaude Opus 5 (Max) | 57.3maxClaude Opus 5 (Max) | 56.6maxClaude Opus 5 (Max) | 53.3maxClaude Opus 5 (Max) |
| 8 | GPT-6 Sol | 54.4maxGPT-6 Sol (Max) | 48.6maxGPT-6 Sol (Max) | 48.6maxGPT-6 Sol (Max) | 51.1maxGPT-6 Sol (Max) | 52.5maxGPT-6 Sol (Max) | 43.6maxGPT-6 Sol (Max) |
| 9 | GPT-6 Luna | 45.2maxGPT-6 Luna (Max) | 37.1maxGPT-6 Luna (Max) | 40.3maxGPT-6 Luna (Max) | 40.3maxGPT-6 Luna (Max) | 45.7maxGPT-6 Luna (Max) | 37.4maxGPT-6 Luna (Max) |
| 10 | Claude Haiku 5.5 | 51.2maxClaude Haiku 5.5 (Max) | 44.9maxClaude Haiku 5.5 (Max) | 43.9maxClaude Haiku 5.5 (Max) | 42.3maxClaude Haiku 5.5 (Max) | 40.4maxClaude Haiku 5.5 (Max) | 44.1maxClaude Haiku 5.5 (Max) |
| 11 | Claude Mythos Preview | — | — | — | — | — | — |
Artificial Analysis 指數
Artificial Analysis Economics Index · Source date unknown
Artificial Analysis Engineering Index · Source date unknown
Artificial Analysis Finance And Accounting Index · Source date unknown
Artificial Analysis Legal Index · Source date unknown
Artificial Analysis Strategy And Ops Index · Source date unknown
Artificial Analysis Healthcare And Medical Index · Source date unknown
PandaNpc 更新 2026年10月21日 06:01 [UTC]
來源: Epoch AI, CC BY 4.0 · Vals.ai · Artificial Analysis (artificialanalysis.ai)
欄位說明
- 編號
- 僅供顯示排序。排序會改變此數字；此並非已發佈的能力排名。
- 模型
- 受評估的模型名稱及其供應商圖示。分數欄中的破折號表示此處沒有已發佈的分數，而非零分。
- Artificial Analysis Economics Index
- 此來源發佈了一項基準分數。請開啟其方法說明以了解任務定義及評分規則。 評估方法 ↗
- Artificial Analysis Engineering Index
- 此來源發佈了一項基準分數。請開啟其方法說明以了解任務定義及評分規則。 評估方法 ↗
- Artificial Analysis Finance And Accounting Index
- 此來源發佈了一項基準分數。請開啟其方法說明以了解任務定義及評分規則。 評估方法 ↗
- Artificial Analysis Legal Index
- 此來源發佈了一項基準分數。請開啟其方法說明以了解任務定義及評分規則。 評估方法 ↗
- Artificial Analysis Strategy And Ops Index
- 此來源發佈了一項基準分數。請開啟其方法說明以了解任務定義及評分規則。 評估方法 ↗
- Artificial Analysis Healthcare And Medical Index
- 此來源發佈了一項基準分數。請開啟其方法說明以了解任務定義及評分規則。 評估方法 ↗
請在同一基準、來源及版本內比較分數。數值越高代表在該報告指標上的表現越強，但不同欄位並不共用同一尺度。