Fiction.LiveBench
ReportedNo description yet: we only describe an eval set when we can do it correctly.
- Kind
- Published by a third party's leaderboard, collected by Epoch AI. The leaderboard is named and linked.
- Link
- https://fiction.live/stories/Fiction-liveBench-Mar-14-2025/oQdzQvKHw8JyXbN87/home
- Score shown
- Epoch's column "16k token score", as a percentage, higher is better; random guessing scores about 0.0%. The source gives no error bars.
- Results
- 61 in all: 26 for open-weight models, 33 for closed ones, 2 where Epoch does not say. 25 linked to a model page here.
Best open-weight result: moonshotai/Kimi-K2.5, 86.1%, number 7 of 61 results.
| # | Model | Weights | Setting | Score | Date | Source |
|---|---|---|---|---|---|---|
| 1 | gpt-5-2025-08-07_mediumOpenAI | closed | effort medium | 97.2% | not dated | Fiction.live leaderboard |
| 2 | o3-pro-2025-06-10_mediumOpenAI | closed | effort medium | 97.2% | not dated | Fiction.live leaderboard |
| 3 | grok-4-0709xAI | closed | default | 94.4% | not dated | Fiction.live leaderboard |
| 4 | grok-4-fastxAI | closed | default | 94.4% | not dated | Fiction.live leaderboard |
| 5 | gemini-2.5-pro-preview-06-05Google DeepMind | closed | default | 91.7% | not dated | Fiction.live leaderboard |
| 6 | o3-2025-04-16_mediumOpenAI | closed | effort medium | 88.9% | not dated | Fiction.live leaderboard |
| 7 | moonshotai/Kimi-K2.5Epoch: kimi-k2.5, Moonshot | open | default | 86.1% | not dated | Fiction.live leaderboard |
| 8 | Qwen/QwQ-32BEpoch: chutes/QwQ-32B, Alibaba | open | default | 83.3% | not dated | Fiction.live leaderboard |
| 9 | claude-3-7-sonnet-20250219_8KAnthropic | closed | default | 83.3% | not dated | Fiction.live leaderboard |
| 10 | deepseek-ai/DeepSeek-V3.2-ExpEpoch: DeepSeek-V3.2-Exp_high, DeepSeek | open | effort high | 83.3% | not dated | Fiction.live leaderboard |
| 11 | o1-2024-12-17_mediumOpenAI | closed | effort medium | 83.3% | not dated | Fiction.live leaderboard |
| 12 | gemini-2.5-flash-preview-05-20Google DeepMind | closed | default | 77.8% | not dated | Fiction.live leaderboard |
| 13 | o4-mini-2025-04-16_mediumOpenAI | closed | effort medium | 77.8% | not dated | Fiction.live leaderboard |
| 14 | deepseek-ai/DeepSeek-R1-0528Epoch: chutes/DeepSeek-R1-0528, DeepSeek | open | default | 75.0% | not dated | Fiction.live leaderboard |
| 15 | Qwen/Qwen3-235B-A22B-Thinking-2507Epoch: chutes/Qwen3-235B-A22B-Thinking-2507, Alibaba | open | default | 75.0% | not dated | Fiction.live leaderboard |
| 16 | Qwen/Qwen3-32BEpoch: chutes/Qwen3-32B, Alibaba | open | default | 74.2% | not dated | Fiction.live leaderboard |
| 17 | deepseek-ai/DeepSeek-R1Epoch: DeepSeek-R1, DeepSeek | open | default | 69.4% | not dated | Fiction.live leaderboard |
| 18 | gpt-5-mini-2025-08-07_mediumOpenAI | closed | effort medium | 69.4% | not dated | Fiction.live leaderboard |
| 19 | MiniMaxAI/MiniMax-M1-80kEpoch: MiniMax-M1-80k, MiniMax | open | default | 69.4% | not dated | Fiction.live leaderboard |
| 20 | Qwen/Qwen3-235B-A22BEpoch: chutes/Qwen3-235B-A22B, Alibaba | open | default | 67.7% | not dated | Fiction.live leaderboard |
| 21 | chatgpt-4o-01-29-2025OpenAI | closed | default | 66.7% | not dated | Fiction.live leaderboard |
| 22 | moonshotai/Kimi-K2-Instruct-0905Epoch: fireworks/Kimi-K2-Instruct-0905, Moonshot | open | default | 66.7% | not dated | Fiction.live leaderboard |
| 23 | gemini-2.5-pro-exp-03-25Google DeepMind | closed | default | 66.7% | not dated | Fiction.live leaderboard |
| 24 | gemini-2.5-pro-preview-03-25Google DeepMind | closed | default | 66.7% | not dated | Fiction.live leaderboard |
| 25 | gemini-2.5-pro-preview-05-06Google DeepMind | closed | default | 66.7% | not dated | Fiction.live leaderboard |
| 26 | grok-3-mini-beta_mediumxAI | closed | effort medium | 66.7% | not dated | Fiction.live leaderboard |
| 27 | qwen-max-2025-01-25Alibaba | closed | default | 66.7% | not dated | Fiction.live leaderboard |
| 28 | qwen3-max-2025-09-23Alibaba | closed | default | 66.7% | not dated | Fiction.live leaderboard |
| 29 | gpt-4.1-2025-04-14OpenAI | closed | default | 63.9% | not dated | Fiction.live leaderboard |
| 30 | gpt-4.5-preview-2025-02-27OpenAI | closed | default | 63.9% | not dated | Fiction.live leaderboard |
| 31 | Qwen/Qwen3-14BEpoch: chutes/Qwen3-14B, Alibaba | open | default | 62.5% | not dated | Fiction.live leaderboard |
| 32 | Qwen/Qwen3-8BEpoch: chutes/Qwen3-8B, Alibaba | open | default | 62.1% | not dated | Fiction.live leaderboard |
| 33 | claude-opus-4-20250514Anthropic | closed | default | 61.1% | not dated | Fiction.live leaderboard |
| 34 | gemini-2.0-flash-001Google DeepMind,Google | closed | default | 61.1% | not dated | Fiction.live leaderboard |
| 35 | moonshotai/Kimi-K2-InstructEpoch: Kimi-K2-Instruct, Moonshot | open | default | 61.1% | not dated | Fiction.live leaderboard |
| 36 | chutes/GLM-4.5-FP8 | not stated | default | 58.3% | not dated | Fiction.live leaderboard |
| 37 | grok-3-betaxAI | closed | default | 58.3% | not dated | Fiction.live leaderboard |
| 38 | deepinfra/Qwen3-Next-80B-A3B-Instruct | not stated | default | 55.6% | not dated | Fiction.live leaderboard |
| 39 | Qwen/Qwen3-235B-A22B-Instruct-2507Epoch: parasail-qwen3-235b-a22b-instruct-2507, Alibaba | open | default | 52.9% | not dated | Fiction.live leaderboard |
| 40 | deepseek-ai/DeepSeek-V3.1Epoch: DeepSeek-V3.1, DeepSeek | open | default | 52.8% | not dated | Fiction.live leaderboard |
| 41 | deepseek-ai/DeepSeek-V3.2-ExpEpoch: DeepSeek-V3.2-Exp, DeepSeek | open | default | 52.8% | not dated | Fiction.live leaderboard |
| 42 | gemini-2.0-flash-thinking-exp-01-21Google DeepMind,Google | closed | default | 52.8% | not dated | Fiction.live leaderboard |
| 43 | deepseek-ai/DeepSeek-V3-0324Epoch: chutes/DeepSeek-V3-0324, DeepSeek | open | default | 50.0% | not dated | Fiction.live leaderboard |
| 44 | claude-3-7-sonnet-20250219Anthropic | closed | default | 50.0% | not dated | Fiction.live leaderboard |
| 45 | o3-mini-2025-01-31_mediumOpenAI | closed | effort medium | 50.0% | not dated | Fiction.live leaderboard |
| 46 | gemini-2.5-flash-lite-preview-06-17-thinkingGoogle DeepMind | closed | default | 47.2% | not dated | Fiction.live leaderboard |
| 47 | gemini-2.5-flash-preview-04-17Google DeepMind | closed | default | 47.2% | not dated | Fiction.live leaderboard |
| 48 | claude-sonnet-4-20250514Anthropic | closed | default | 46.9% | not dated | Fiction.live leaderboard |
| 49 | meta-llama/Llama-4-Maverick-17B-128E-InstructEpoch: chutes/Llama-4-Maverick-17B-128E-Instruct, Meta AI | open | default | 46.2% | not dated | Fiction.live leaderboard |
| 50 | openai/gpt-oss-120bEpoch: chutes/gpt-oss-120b_high, OpenAI | open | effort high | 44.4% | not dated | Fiction.live leaderboard |
| 51 | gpt-4.1-mini-2025-04-14OpenAI | closed | default | 44.4% | not dated | Fiction.live leaderboard |
| 52 | gpt-5-nano-2025-08-07_mediumOpenAI | closed | effort medium | 44.4% | not dated | Fiction.live leaderboard |
| 53 | Qwen/Qwen3-Next-80B-A3B-InstructEpoch: chutes/Qwen3-Next-80B-A3B-Instruct | open | default | 41.7% | not dated | Fiction.live leaderboard |
| 54 | gemini-2.0-pro-exp-02-05Google DeepMind | closed | default | 41.7% | not dated | Fiction.live leaderboard |
| 55 | Qwen/Qwen3-30B-A3BEpoch: chutes/Qwen3-30B-A3B, Alibaba | open | default | 40.6% | not dated | Fiction.live leaderboard |
| 56 | openai/gpt-oss-120bEpoch: chutes/gpt-oss-120b, OpenAI | open | default | 36.1% | not dated | Fiction.live leaderboard |
| 57 | chutes/Llama-4-Scout-17B-16E InstructMeta AIno page match | open | default | 36.0% | not dated | Fiction.live leaderboard |
| 58 | google/gemma-3-27b-itEpoch: chutes/Gemma-3-27b-It, Google DeepMind | open | default | 33.3% | not dated | Fiction.live leaderboard |
| 59 | meta-llama/Llama-3.3-70B-InstructEpoch: Llama-3.3-70B-Instruct, Meta AI | open | default | 33.3% | not dated | Fiction.live leaderboard |
| 60 | gpt-4.1-nano-2025-04-14OpenAI | closed | default | 25.0% | not dated | Fiction.live leaderboard |
| 61 | nvidia/NVIDIA-Nemotron-Nano-9B-v2Epoch: nvidia-nemotron-nano-9b-v2 | open | default | 25.0% | not dated | Fiction.live leaderboard |
Open-weight rows are shaded. Dates are when the result was run or evaluated, where the data says. Data from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Epoch AI. Method