Fiction.LiveBench 120k
Long-context comprehension of fiction at 120k tokens.
- Published by
- Fiction.livedata via Epoch AI Benchmarking Hub
- Category
- Long context
- Index weight
- 1.0
- Models
- 35
- Data as of
- 9 Sept 2026
CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.
- 1o3100.0%
- 2GPT-596.9%
- 3Grok 496.9%
- 4Gemini 2.5 Pro90.6%
- 5o3-pro88.9%
- 6Kimi K2.578.1%
- 7Grok 4 Fast75.0%
- 8Qwen3 235B A22B68.8%
- 9Gemini 2.5 Flash68.8%
- 10Chatgpt 4o 01.2965.6%
- 11Grok 3 mini65.6%
- 12GPT-4.563.9%
- 13GPT-5 mini62.5%
- 14o4-mini62.5%
- 15Qwen3-Next 80B A3B62.5%
35 of 35
| # | |||||
|---|---|---|---|---|---|
| 1 |  o3OpenAI | 100.0% | – | 53.6 | – |
| 2 | GPT-5OpenAI | 96.9% | – | 58.5 | – |
| 3 |  Grok 4xAI | 96.9% | 0709 | 57.2 | – |
| 4 |  Gemini 2.5 ProGoogle | 90.6% | – | 54.1 | – |
| 5 | o3-proOpenAI | 88.9% | – | – | – |
| 6 |  Kimi K2.5Moonshot AIopen | 78.1% | kimi-k2.5 | 56.1 | – |
| 7 | Grok 4 FastxAI | 75.0% | – | 46.5 | – |
| 8 |  Qwen3 235B A22BAlibabaopen | 68.8% | thinking, 2507 | 51.7 | – |
| 9 | Gemini 2.5 FlashGoogle | 68.8% | – | 48.8 | – |
| 10 | Chatgpt 4o 01.29OpenAI | 65.6% | 2025 | – | – |
| 11 | Grok 3 minixAI | 65.6% | – | – | – |
| 12 | GPT-4.5OpenAI | 63.9% | – | 49.6 | – |
| 13 | GPT-5 miniOpenAI | 62.5% | – | 54.5 | – |
| 14 | o4-miniOpenAI | 62.5% | – | 52.4 | – |
| 15 | Qwen3-Next 80B A3BAlibabaopen | 62.5% | – | 48.4 | – |
| 16 | GPT-4.1OpenAI | 62.5% | – | 47.3 | – |
| 17 | Gemini 2.0 FlashGoogle | 62.5% | – | 42.3 | – |
| 18 |  MiniMax-M1MiniMaxopen | 59.4% | – | 48.2 | – |
| 19 | Grok 3xAI | 58.3% | – | 50.0 | – |
| 20 |  Claude 3.7 SonnetAnthropic | 53.1% | – | 49.8 | – |
| 21 | o1OpenAI | 53.1% | – | 48.3 | – |
| 22 | GPT-4.1 miniOpenAI | 46.9% | – | 44.5 | – |
| 23 | Parasail Qwen3 235B A22bAlibaba | 44.4% | 2507 | – | – |
| 24 | o3-miniOpenAI | 43.8% | – | 45.6 | – |
| 25 | Kimi K2Moonshot AIopen | 40.6% | 0905 | 48.8 | – |
| 26 | Claude Opus 4Anthropic | 37.5% | – | 50.7 | – |
| 27 | Gemini 2.0 ProGoogle | 37.5% | – | 47.0 | – |
| 28 | Gemini 2.0 FlashGoogle | 37.5% | – | 45.1 | – |
| 29 | Claude Sonnet 4Anthropic | 36.4% | – | 49.5 | – |
| 30 |  Llama 4 MaverickMetaopen | 36.4% | – | 42.6 | – |
| 31 |  DeepSeek R1DeepSeekopen | 33.3% | DeepSeek-R1 | 48.9 | – |
| 32 | Llama 4 ScoutMetaopen | 27.3% | – | 39.7 | – |
| 33 | Gemini 2.5 Flash-LiteGoogle | 21.9% | – | 46.7 | – |
| 34 | GPT-5 nanoOpenAI | 21.9% | – | 46.5 | – |
| 35 | GPT-4.1 nanoOpenAI | 18.8% | – | 38.1 | – |