HyperAI
Command Palette
Search for a command to run...
AA-LCR v1.1: Long Context Reasoning
一项高难度基准，衡量语言模型从 10k 至 100k token 的长文档（使用 cl100k_base 分词器计量）中提取、推理并综合信息的能力。
- 模型
- 259 当前，596 总共
- 中位数
- 61.3%
- 范围
- 0.0%–88.3%
| 排名 | 比较 | 模型 | 评分 |
|---|---|---|---|
| 151 | 49.0% | ||
| 152 |  Qwen3 Coder Next Alibaba | 47.0% | |
| 153 | 46.7% | ||
| 154 |  gpt-oss-120b (Low) OpenAI | 46.0% | |
| 155 | Qwen3.5 9B (Non-reasoning) Alibaba | 46.0% | |
| 156 | 45.0% | ||
| 157 | 42.3% | ||
| 158 |  G9v3-3B AI9Stars | 41.3% | |
| 159 | 41.0% | ||
| 160 |  Solar Open 100B (Reasoning) Upstage | 41.0% | |
| 161 | 39.7% | ||
| 162 |  INTELLECT-3 (Based on GLM-4.5-Air) Prime Intellect | 39.7% | |
| 163 | 39.7% | ||
| 164 | 39.3% | ||
| 165 |  Celeris-1 Celeris | 38.0% | |
| 166 | 38.0% | ||
| 167 |  Trinity Large Thinking Arcee AI | 38.0% | |
| 168 |  North Mini Code Cohere | 37.3% | |
| 169 |  HyperNova 60B 2605 (High, Based on gpt-oss-120b) Multiverse Computing | 36.3% | |
| 170 | 35.0% | ||
| 171 |  K2-V2 (High) Institute of Foundation Models | 35.0% | |
| 172 | gpt-oss-20b (High) OpenAI | 34.7% | |
| 173 | Qwen3.5 4B (Non-reasoning) Alibaba | 34.7% | |
| 174 |  Nova Premier Amazon | 34.3% | |
| 175 |  Gemma 4 E4B (Reasoning) Google | 32.0% | |
| 176 | gpt-oss-20b (Low) OpenAI | 31.0% | |
| 177 | 30.0% | ||
| 178 |  LongCat Flash Lite LongCat | 28.3% | |
| 179 | 28.3% | ||
| 180 | K2-V2 (Medium) Institute of Foundation Models | 28.0% | |
| 181 | 28.0% | ||
| 182 |  Llama 4 Scout Meta | 27.7% | |
| 183 |  Ministral 3 14B Mistral | 26.3% | |
| 184 | Ministral 3 8B Mistral | 25.7% | |
| 185 | 25.3% | ||
| 186 | 24.7% | ||
| 187 | 24.3% | ||
| 188 | 24.0% | ||
| 189 | 24.0% | ||
| 190 |  Cogito v2.1 (Reasoning) Deep Cogito | 22.7% | |
| 191 | 22.7% | ||
| 192 |  Hermes 4 - Llama-3.1 405B (Reasoning) Nous Research | 22.3% | |
| 193 | Hermes 4 - Llama-3.1 405B (Non-reasoning) Nous Research | 22.0% | |
| 194 | 21.7% | ||
| 195 | Command A Cohere | 21.3% | |
| 196 | Qwen3.5 2B (Reasoning) Alibaba | 21.0% | |
| 197 |  Ring-flash-2.0 InclusionAI | 21.0% | |
| 198 | 20.7% | ||
| 199 | K2-V2 (Low) Institute of Foundation Models | 20.3% | |
| 200 | DiffusionGemma 26B A4B Google | 19.7% |
