Fiction.LiveBench
Rank #24 of 51 · 45 models scored · top score 97.2
unknown
Strongest on discrimination, weakest on contamination resistance. Caveat: the public test set leaves contamination on the table.
Weighted composite. Each signal is computed from the source-backed leaderboard below.
Does it still separate models?
How far from ceiling / clustered at the top?
Public vs held-out; training-leak risk.
Apples-to-apples, or mixed / vendor-optimized?
How old is the benchmark?
| # | Model | Score | Evidence | Measured |
|---|---|---|---|---|
| 1 | o3-pro | 97.2 | unverified | 2025-06-10 |
| 2 | GPT-5 | 97.2 | unverified | 2025-08-07 |
| 3 | Grok 4 | 94.4 | unverified | 2025-07-09 |
| 4 | Grok 4 Fast | 94.4 | unverified | 2025-09-19 |
| 5 | Gemini 2.5 Pro (Jun 2025) | 91.7 | unverified | 2025-06-05 |
| 6 | o3 | 88.9 | unverified | 2024-12-20 |
| 7 | Kimi K2.5 | 86.1 | unverified | 2026-02-02 |
| 8 | Claude 3.7 Sonnet | 83.3 | unverified | 2025-02-24 |
| 9 | o1 | 83.3 | unverified | 2024-12-05 |
| 10 | DeepSeek-V3.2-Exp | 83.3 | unverified | 2025-09-29 |
| 11 | o4-mini | 77.8 | unverified | 2025-04-16 |
| 12 | Gemini 2.5 Flash (May 2025) | 77.8 | unverified | 2025-05-20 |
| 13 | Qwen3-235B-A22B-Thinking (Jul 2025) | 75 | unverified | 2025-07-25 |
| 14 | DeepSeek-R1 (May 2025) | 75 | unverified | 2025-05-28 |
| 15 | DeepSeek-R1 | 69.4 | unverified | 2025-01-20 |
| 16 | GPT-5 mini | 69.4 | unverified | 2025-08-07 |
| 17 | Qwen3-235B-A22B | 67.7 | unverified | 2025-04-28 |
| 18 | Grok-3 mini | 66.7 | unverified | 2025-02-19 |
| 19 | Kimi K2 (Sep 2025) | 66.7 | unverified | 2025-07-11 |
| 20 | Qwen2.5-Max | 66.7 | unverified | 2025-01-25 |
| 21 | Qwen3-Max | 66.7 | unverified | 2025-09-05 |
| 22 | Gemini 2.5 Pro (Mar 2025) | 66.7 | unverified | 2025-03-25 |
| 23 | Gemini 2.5 Pro (May 2025) | 66.7 | unverified | 2025-05-06 |
| 24 | GPT-4.1 | 63.9 | unverified | 2025-04-14 |
| 25 | GPT-4.5 | 63.9 | unverified | 2025-02-27 |
| 26 | Kimi K2 (Jul 2025) | 61.1 | unverified | 2025-07-11 |
| 27 | Claude Opus 4 | 61.1 | unverified | 2025-05-22 |
| 28 | Gemini 2.0 Flash (Feb 2025) | 61.1 | unverified | 2024-12-11 |
| 29 | Grok 3 | 58.3 | unverified | 2025-02-17 |
| 30 | Qwen3-235B-A22B-Instruct (Jul 2025) | 52.9 | unverified | 2025-07-25 |
| 31 | DeepSeek-V3.1 | 52.8 | unverified | 2025-08-21 |
| 32 | Gemini 2.0 Flash Thinking (Jan 2025) | 52.8 | unverified | 2025-01-21 |
| 33 | o3-mini | 50 | unverified | 2025-01-31 |
| 34 | DeepSeek-V3 (Mar 2025) | 50 | unverified | 2025-03-24 |
| 35 | Gemini 2.5 Flash (Apr 2025) | 47.2 | unverified | 2025-04-17 |
| 36 | Claude Sonnet 4 | 46.9 | unverified | 2025-05-22 |
| 37 | Llama 4 Maverick | 46.2 | unverified | 2025-04-05 |
| 38 | gpt-oss-120b | 44.4 | unverified | 2025-08-05 |
| 39 | GPT-4.1 mini | 44.4 | unverified | 2025-04-14 |
| 40 | GPT-5 nano | 44.4 | unverified | 2025-08-07 |
| 41 | Gemini 2.0 Pro | 41.7 | unverified | 2024-12-11 |
| 42 | Llama 4 Scout | 36 | unverified | 2025-04-05 |
| 43 | Gemma 3 27B | 33.3 | unverified | 2025-03-12 |
| 44 | Llama 3.3 70B | 33.3 | unverified | 2024-12-06 |
| 45 | GPT-4.1 nano | 25 | unverified | 2025-04-14 |