HyperAI
Command Palette
Search for a command to run...
Terminal-Bench-Science 0.1
一项包含 70 个任务的基准，覆盖科研工作流，由生命、物理、地球、数学和工程科学领域的专家编写并评审，每个任务都在终端中完成，并由各自的测试集校验。
- 智能指数权重
- 15%
- 模型
- 109 当前，135 总共
- 中位数
- 7.6%
- 范围
- 0.0%–73.3%
| 排名 | 比较 | 模型 | 评分 | 成本 / 任务 | 时间 / 任务 | 输出 token / 任务 |
|---|---|---|---|---|---|---|
| 51 |  GLM-5.3 (Max) Z AI | 9.5% | $14.91 | 64m 59s | 268K | |
| 52 |  Grok 4.7 (High) SpaceXAI | 9.5% | $16.85 | 40m 55s | 165K | |
| 53 |  GPT-6 Luna (Max) OpenAI | 8.6% | $0.45 | 23m 6s | 152K | |
| 54 |  Gemini 3.8 Flash (Low) Google | 8.1% | — | — | — | |
| 55 | GPT-6.1 Terra (Low) OpenAI | 7.6% | $0.12 | 2m 2s | 8.27K | |
| 56 |  Claude Haiku 5.5 (Medium) Anthropic | 7.1% | $0.25 | 8m 2s | 109K | |
| 57 |  DeepSeek V4 Pro 0813 (Max) DeepSeek | 5.7% | $5.63 | 50m 7s | 208K | |
| 58 |  Inkling 2 Small Thinking Machines | 5.7% | $3.68 | 48m 19s | 370K | |
| 59 |  MiMo-V2.6-Pro Xiaomi | 5.7% | $0.63 | 85m 44s | 290K | |
| 60 | GPT-6 Luna (Xhigh) OpenAI | 5.2% | $0.34 | 15m 20s | 94.8K | |
| 61 | GLM-5.3-Flash Z AI | 4.8% | $1.60 | 91m 8s | 271K | |
| 62 | Grok 4.7 (Low) SpaceXAI | 4.8% | $8.88 | 20m 59s | 81.3K | |
| 63 |  Ling 3.1 Flash 1019 InclusionAI | 4.3% | $10.29 | — | 539K | |
| 64 | MiMo-V2.6-Flash Xiaomi | 4.3% | $0.38 | 145m 49s | 307K | |
| 65 |  Mistral Large 4 (High) Mistral | 4.3% | $8.10 | 29m 43s | 209K | |
| 66 |  Agnes 2.5 Pro Theta Sapiens AI | 3.8% | $1.16 | 49m 11s | 570K | |
| 67 | GLM-5.3 (Low) Z AI | 3.8% | $3.32 | 33m 59s | 120K | |
| 68 |  MiniMax-M3.1-Flash MiniMax | 3.3% | $1.50 | 31m 29s | 302K | |
| 69 | Mistral Large 4 (Medium) Mistral | 3.3% | $5.78 | 22m 23s | 151K | |
| 70 | GPT-6 Luna (High) OpenAI | 2.9% | $0.24 | 12m 52s | 74.8K | |
| 71 | 2.4% | $0.77 | 17m 8s | 129K | ||
| 72 |  Step 5 Preview StepFun | 2.4% | $7.28 | 34m 10s | 296K | |
| 73 | Claude Haiku 5.5 (Low) Anthropic | 1.9% | $0.11 | 4m 27s | 58.5K | |
| 74 | 1.9% | $5.22 | 14m 30s | 150K | ||
| 75 | GPT-6 Luna (Medium) OpenAI | 1.4% | $0.13 | 8m 5s | 48.3K | |
| 76 | GPT-6 Luna (Low) OpenAI | 1.0% | $0.0078 | 1m 38s | 8.46K | |
| 77 | 1.0% | $0.28 | 6m 4s | 32.8K | ||
| 78 |  Nanbeige4.3-33B Nanbeige | 1.0% | — | — | 325K | |
| 79 |  Qwen3.8 27B (Medium) Alibaba | 1.0% | $2.57 | 51m 11s | 137K | |
| 80 |  Apodex 1.1 Apodex | 0.5% | $1.17 | 21m 33s | 184K | |
| 81 | MiniMax-M3 MiniMax | 0.5% | $4.01 | 25m 56s | 193K | |
| 82 | Qwen3.8 27B (Xhigh) Alibaba | 0.5% | $12.21 | 104m 20s | 289K | |
| 83 | Qwen3.8 27B (Low) Alibaba | 0.5% | $2.51 | 53m 23s | 161K | |
| 84 |  Beam 1.0 Reflection | 0.5% | $0.87 | 3m 46s | 65.1K | |
| 85 |  Command A+ Cohere | 0.0% | $0 | 4m 58s | 59.3K | |
| 86 | 0.0% | $2.72 | 44m 58s | 179K | ||
| 87 | Gemini 3.5 Flash-Lite Google | 0.0% | $1.06 | 2m 30s | 46.6K | |
| 88 | Gemma 4 31B (Reasoning) Google | 0.0% | — | — | — | |
| 89 | gpt-oss-120b (High) OpenAI | 0.0% | $0.42 | 10m 9s | 98.4K | |
| 90 | 0.0% | — | — | — | ||
| 91 | 0.0% | — | — | 57.2K | ||
| 92 | 0.0% | $0.06 | 15m 39s | 90.6K | ||
| 93 | Inkling (Xhigh) Thinking Machines | 0.0% | $2.02 | 14m 22s | 83.9K | |
| 94 | Inkling Small Thinking Machines | 0.0% | $0.38 | 14m 26s | 62K | |
| 95 |  K-EXAONE 2.0 LG AI Research | 0.0% | — | — | 100K | |
| 96 |  K2 Horizon 3.7B Institute of Foundation Models | 0.0% | — | — | — | |
| 97 | K2 Horizon 375B A23B Institute of Foundation Models | 0.0% | $0 | 1m 12s | 13.6K | |
| 98 | K2 Horizon 7B Institute of Foundation Models | 0.0% | — | — | — | |
| 99 | K2 Horizon MoVA 36B A4B Institute of Foundation Models | 0.0% | — | — | — | |
| 100 | Mistral Medium 3.5 Mistral | 0.0% | $2.38 | 9m 58s | 83.8K |
