Fiction.LiveBench - 120K tokens
Chat & Writing · 2025-02-19
Fiction.LiveBench quiz accuracy when the story is cut to about 120K tokens, so the relevant passages sit inside that much surrounding fiction. The same private, open-ended comprehension questions are asked at every length; question counts are not published.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 3 Flash Preview | 100.0 |
| GPT-5.2 | 100.0 |
| o3 | 100.0 |
| Gemini 3 Pro Preview | 96.9 |
| GPT-5 | 96.9 |
| Grok 4 | 96.9 |
| Opus 4.5 | 93.8 |
| Opus 4.6 | 93.8 |
| Gemini 2.5 Pro | 87.5 |
| GLM-5 | 85.7 |
| Sonoma Sky Alpha | 84.4 |
| Sonnet 4 | 81.3 |
| Kimi K2.5 | 78.1 |
| Qwen3.5 Plus (2026-02-15) | 76.2 |
| Sonnet 4.5 | 75.0 |
| GPT-5.2 Pro | 75.0 |
| Grok 4 Fast | 75.0 |
| Qwen3.6 Plus (2026-04-02) | 75.0 |
| Qwen3 235B A22B Thinking 2507 | 68.8 |
| ChatGPT-4o Latest (source-unspecified snapshot) | 65.6 |
| Gemini 2.5 Flash | 65.6 |
| GPT-4.5 | 63.9 |
| Opus 4 | 62.5 |
| DeepSeek-V3.1 | 62.5 |
| GPT-4.1 | 62.5 |
| GPT-5 Mini | 62.5 |
| O4 Mini | 62.5 |
| Qwen3-Next-80B-A3B-Thinking | 62.5 |
| MiniMax M1 | 59.4 |
| Gemini 2.5 Flash Preview 04-17 | 53.1 |
| Qwen3 30B A3B Thinking 2507 | 53.1 |
| Nemotron 3 Nano 30B A3B | 50.0 |
| Qwen3 Next 80B A3B Instruct | 46.9 |
| Qwen3 235B A22B Instruct 2507 | 44.4 |
| GPT-5 Chat | 43.8 |
| Kimi K2 Instruct 0905 | 40.6 |
| MiniMax M2.5 | 40.6 |
| Llama 4 Maverick Instruct | 36.4 |
| Gemini 2.5 Flash-Lite | 31.3 |
| Llama 4 Scout Instruct | 27.3 |