Write N of X

Write exactly N words/sentences/paragraphs...

20 word summary

0-shot Rule following
Model Run 1 Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10 Total
GPT-4o Mini (temp=0)100%100%100%100%100%100%100%100%100%100%100%
Gemini 3 Pro (Preview)100%100%100%100%100%100%100%100%100%100%100%
Z.AI GLM 4.7100%100%100%100%100%100%100%100%100%100%100%
o4 Mini100%100%100%100%100%100%100%100%100%100%100%
Z.AI GLM 4.7 Flash100%100%100%100%100%100%100%100%100%100%100%
MoonshotAI: Kimi K2.5100%100%100%100%100%100%100%100%100%100%100%
Claude 3.5 Sonnet (new)100%100%100%100%100%100%100%100%100%100%100%
GPT-4o Mini (temp=1)100%100%100%100%100%100%100%100%100%100%100%
GPT-4o, May 13th (temp=0)100%100%100%100%100%100%100%100%100%100%100%
o4 Mini High100%100%100%100%100%100%100%100%100%98%100%
Llama 3 TenyxChat-DaybreakStorywriter 70B100%100%100%100%100%100%100%100%100%98%100%
GPT-4o, May 13th (temp=1)100%100%100%100%100%100%100%100%100%98%100%
Gemini 2.5 Pro100%100%100%100%100%100%100%100%100%98%100%
GPT-4.1100%100%100%100%100%100%100%100%100%98%100%
Gemini 3 Flash (Preview)100%100%100%100%100%100%100%100%100%98%100%
GPT-4.1 Nano100%100%100%100%100%100%100%100%100%98%100%
GPT-4.1 Mini100%100%100%100%100%100%100%100%98%98%100%
Claude 3.7 Sonnet100%100%100%100%100%100%100%100%98%98%100%
Llama 3.1 405B100%100%100%100%100%100%100%100%98%98%100%
GPT-4o, Aug. 6th (temp=1)100%100%100%100%100%100%100%100%98%98%100%
Z.AI GLM 4.5100%100%100%100%100%100%100%98%98%98%99%
GPT-4 Turbo100%100%100%100%100%100%100%98%98%98%99%
Llama 3 70B100%100%100%100%100%100%100%98%98%98%99%
Z.AI GLM 4.6100%100%100%100%100%100%98%98%98%98%99%
Claude 3.5 Haiku100%100%100%100%100%98%98%98%98%98%99%
Claude Opus 4100%100%100%100%100%98%98%98%98%98%99%
GPT-4o, Aug. 6th (temp=0)98%98%98%98%98%98%98%98%98%98%98%
Claude Sonnet 4100%100%100%98%98%98%98%98%98%92%98%
Gemini 2.5 Flash Lite100%100%100%100%98%98%98%98%92%92%98%
Claude 3.5 Sonnet100%100%100%100%100%100%98%98%98%77%97%
Gemini Pro 1.5100%100%100%100%100%98%98%98%98%77%97%
DeepSeek-V2 Chat100%100%100%100%100%100%100%100%92%77%97%
Llama 3.2 11B (Vision)100%100%100%100%100%100%100%98%92%77%97%
Qwen 2.5 72B100%100%100%100%100%100%100%100%92%54%95%
Claude Opus 4.5100%100%100%100%98%98%98%98%92%54%94%
Cohere Command R+ (Aug. 2024)100%100%98%98%98%98%92%77%77%77%92%
Claude Opus 4.698%98%98%98%92%92%92%92%92%54%91%
Llama 3.1 8B100%100%98%98%98%92%77%77%77%77%90%
Hermes 3 405B100%100%100%100%100%100%98%98%92%0%89%
Claude Sonnet 4.5100%98%98%92%92%92%92%77%77%54%87%
Mistral Small Creative98%98%92%92%92%92%92%77%77%54%87%
Gemma 2 27B100%100%100%100%100%98%92%54%54%54%85%
Gemma 2 9B100%100%100%100%100%100%92%54%54%27%83%
Llama 3.2 3B100%100%100%100%98%98%92%77%27%0%79%
AI21 Jamba 1.5 Large100%100%100%98%92%77%77%54%9%0%71%
Phi-3.5 Mini 128k100%100%100%92%92%92%54%54%9%0%69%
Llama 3.1 Nemotron 70B100%100%100%100%98%77%54%54%9%0%69%
Claude 3 Haiku100%100%100%100%98%54%54%54%27%2%69%
Claude Haiku 4.5100%100%100%98%77%54%54%27%9%9%63%
AI21 Jamba 1.5 Mini100%100%100%100%98%92%0%0%0%0%59%
Inflection 3 (Productivity)100%100%92%92%92%54%0%0%0%0%53%
Claude 2.198%92%92%77%77%27%27%27%2%0%52%
Gemini 2.5 Flash98%92%92%54%54%54%54%9%9%2%52%
Llama 3.2 1B100%100%92%77%54%54%27%0%0%0%50%
Phi-3 Mini 128k100%100%98%77%54%54%0%0%0%0%48%
Gemini Flash 1.5100%100%98%77%27%27%27%0%0%0%46%
Llama 3.1 Euryale 70B v2.2100%98%92%92%27%9%2%0%0%0%42%
Qwen 2 72B100%100%98%92%0%0%0%0%0%0%39%
Llama 3 Euryale 70B v2.1100%100%92%9%2%0%0%0%0%0%30%
Claude 3.0 Sonnet98%77%77%27%9%2%0%0%0%0%29%
WizardLM 2 8x22b98%92%92%0%0%0%0%0%0%0%28%
Inflection 3 (PI)100%92%27%27%27%2%0%0%0%0%28%
Ministral 3B92%77%54%0%0%0%0%0%0%0%22%
EVA Qwen 2.5 14B92%92%2%0%0%0%0%0%0%0%19%
AI21 Jamba98%77%2%0%0%0%0%0%0%0%18%
Writer: Palmyra X554%54%27%27%9%2%0%0%0%0%17%
Llama 3.2 90B (Vision)54%54%27%27%9%2%0%0%0%0%17%
Mistral Medium54%54%9%9%9%2%2%0%0%0%14%
Lumimaid v0.2 8B92%27%2%0%0%0%0%0%0%0%12%
Hermes 3 70B54%54%0%0%0%0%0%0%0%0%11%
Sao10K L3.1 70B Hanami x198%2%0%0%0%0%0%0%0%0%10%
Ministral 8B100%0%0%0%0%0%0%0%0%0%10%
Phi-3 Medium 128k100%0%0%0%0%0%0%0%0%0%10%
Claude 2.0100%0%0%0%0%0%0%0%0%0%10%
Mistral NeMO92%0%0%0%0%0%0%0%0%0%9%
Rocinante 12B92%0%0%0%0%0%0%0%0%0%9%
Magnum v2 72B92%0%0%0%0%0%0%0%0%0%9%
Llama 3.1 70B77%2%2%0%0%0%0%0%0%0%8%
Magnum 72B77%0%0%0%0%0%0%0%0%0%8%
Cohere Command R+ (Apr. 2024)54%9%2%2%2%2%2%2%2%2%8%
Mistral Large27%9%9%0%0%0%0%0%0%0%5%
Liquid: LFM 40B MoE27%9%0%0%0%0%0%0%0%0%4%
MythoMist 7B2%0%0%0%0%0%0%0%0%0%0%
MN GRAND Gutenberg Lyra4 12B Madness2%0%0%0%0%0%0%0%0%0%0%
Mistral Nemo 12B Celeste0%0%0%0%0%0%0%0%0%0%0%
Goliath 120B0%0%0%0%0%0%0%0%0%0%0%
lzlv 70B0%0%0%0%0%0%0%0%0%0%0%
Mistral Large 20%0%0%0%0%0%0%0%0%0%0%
Qwen 2 7B0%0%0%0%0%0%0%0%0%0%0%
Hermes 2 Theta 8B0%0%0%0%0%0%0%0%0%0%0%
MythoMax 13B0%0%0%0%0%0%0%0%0%0%0%
Toppy M 7B0%0%0%0%0%0%0%0%0%0%0%
Fimbulvetr 11B v20%0%0%0%0%0%0%0%0%0%0%
57.77%