HyperAI
Command Palette
Search for a command to run...
AutomationBench-AA
シミュレートされたSaaSアプリケーション環境におけるエージェントのタスク完了度を測定するベンチマークで、ガードレール違反なしに完了した各タスク目標の割合で採点する。
- インテリジェンス指標の重み
- 5%
- モデル
- 150 現在、 256 合計
- 中央値
- 46.0%
- 範囲
- 0.2%–77.8%
| ランク | 比較する | モデル | スコア | コスト / タスク | 時間 / タスク | 出力 token / タスク |
|---|---|---|---|---|---|---|
| 51 |  Qwen3.8 2.4T A95B Alibaba | 57.2% | — | — | — | |
| 52 |  Grok 4.7 (Low) SpaceXAI | 56.7% | $0.41 | 2m 56s | 11.4K | |
| 53 |  DeepSeek V4 Pro 0813 (Max) DeepSeek | 56.7% | $0.34 | 8m 54s | 37K | |
| 54 |  GPT-6.1 Terra (Low) OpenAI | 56.2% | $0.078 | 1m 23s | 5.66K | |
| 55 | Qwen3.8 Max (0902) Alibaba | 56.2% | $0.40 | 7m 9s | 21.9K | |
| 56 | Qwen3.8-Flash-Next Alibaba | 55.9% | — | — | — | |
| 57 | 54.9% | $0.26 | 1m 28s | 12.2K | ||
| 58 | GPT-6 Luna (Max) OpenAI | 53.2% | $0.036 | 5m 11s | 34K | |
| 59 |  Nanbeige4.3-33B Nanbeige | 53.1% | — | — | 54.4K | |
| 60 | 52.9% | $0.49 | 1m 50s | 11.9K | ||
| 61 |  Mistral Large 4 (Medium) Mistral | 52.8% | $0.12 | 2m 9s | 14.6K | |
| 62 | GPT-6.1 Sol (Low) OpenAI | 52.6% | $0.16 | 1m 37s | 5.46K | |
| 63 |  GLM-5.3 (Low) Z AI | 52.0% | $0.21 | 2m 59s | 10.5K | |
| 64 | 51.0% | $0.088 | 1m 29s | 6.14K | ||
| 65 |  Step 5 Preview StepFun | 51.0% | $0.22 | 3m 26s | 29.7K | |
| 66 | 49.7% | $0.46 | 2m 48s | 29.1K | ||
| 67 | 49.4% | $0.24 | 1m 20s | 11.2K | ||
| 68 |  Agnes 2.5 Pro Theta Sapiens AI | 48.8% | $0.039 | 4m 6s | 47.5K | |
| 69 | 48.8% | $0.044 | 1m 54s | 14.3K | ||
| 70 |  Beam 1.0 Reflection | 48.5% | $0.092 | 1m 21s | 23.5K | |
| 71 | Qwen3.8 27B (Xhigh) Alibaba | 48.2% | $0.28 | 11m 38s | 32.2K | |
| 72 | GPT-6 Luna (High) OpenAI | 47.8% | $0.021 | 2m 52s | 16.7K | |
| 73 | GPT-6 Luna (Xhigh) OpenAI | 47.8% | $0.023 | 3m 14s | 20K | |
| 74 | DeepSeek V4 Flash Vision (Max) DeepSeek | 47.5% | — | — | — | |
| 75 |  JT-4.1 Flash 236B A21B (Reasoning) China Mobile | 46.5% | — | — | — | |
| 76 |  Inkling 2 Small Thinking Machines | 45.6% | $0.058 | 2m 28s | 18.9K | |
| 77 | GPT-6 Luna (Medium) OpenAI | 40.5% | $0.017 | 2m 4s | 12.4K | |
| 78 | Qwen3.8 27B (Medium) Alibaba | 40.3% | $0.25 | 11m 58s | 32K | |
| 79 |  K2 Horizon 375B A23B Institute of Foundation Models | 37.2% | $0 | 2m 47s | 31.4K | |
| 80 |  Gemini 3.8 Flash (Low) Google | 36.5% | — | — | — | |
| 81 | 36.4% | — | — | — | ||
| 82 |  Claude Haiku 5.5 (Xhigh) Anthropic | 36.0% | $0.028 | 1m 58s | 28K | |
| 83 | Gemini 3.1 Pro Preview Google | 35.4% | — | — | — | |
| 84 | Claude Haiku 5.5 (Max) Anthropic | 35.4% | $0.042 | 3m 21s | 51K | |
| 85 | 34.9% | $0.26 | 5m 42s | 22.7K | ||
| 86 | Claude Haiku 5.5 (High) Anthropic | 33.7% | $0.023 | 1m 32s | 21.1K | |
| 87 | Qwen3.8 27B (Low) Alibaba | 30.1% | $0.30 | 11m 54s | 35.8K | |
| 88 | Claude Haiku 5.5 (Medium) Anthropic | 28.6% | $0.019 | 1m 15s | 16.9K | |
| 89 |  Apodex 1.1 Apodex | 28.2% | $0.23 | 5m 57s | 50.8K | |
| 90 | K2 Horizon MoVA 36B A4B Institute of Foundation Models | 26.1% | — | — | — | |
| 91 | Gemini 3.5 Flash-Lite Google | 25.0% | $0.20 | 1m 8s | 21.1K | |
| 92 |  Kimi K2.7 Code Kimi | 24.5% | — | — | — | |
| 93 |  Quasar 438B (Max, Based on GLM-5.2) Multiverse Computing | 23.9% | — | — | — | |
| 94 | Claude Haiku 5.5 (Low) Anthropic | 22.9% | $0.015 | 59.4s | 13K | |
| 95 |  Solar Mini 4 Upstage | 22.3% | — | — | — | |
| 96 |  MiniMax-M3 MiniMax | 21.3% | $0.12 | 2m 47s | 20.8K | |
| 97 | K2 Horizon 7B Institute of Foundation Models | 20.9% | — | — | — | |
| 98 |  Hy3 Tencent | 18.4% | — | — | — | |
| 99 | 17.6% | — | — | — | ||
| 100 | Qwen3.7 Plus Alibaba | 17.4% | — | — | — |
