Reasoning benchmark

WinoGrande leaderboard

As of October 2026, Llama 3.1-405B has the highest published WinoGrande score on Noometry at 89.2%, out of 43 models with results.

Last verified October 10, 2026

About WinoGrande

Pronoun-resolution problems that need commonsense reasoning.

Category: Reasoning
Introduced: 2019
Format: Binary choice
Unit: Percent (random guessing ≈ 50%)
Official site: winogrande.allenai.org

Top 15 models

Llama 3.1-405B 89.2%
Claude 3 Opus 88.5%
GPT-4 87.5%
GPT-4 87.5%
Falcon-180B 87.1%
DeepSeek-V2 (MoE-236B, May 2024) 86.3%
DeepSeek-V3 85.2%
Deepseek Coder v2 83.7%
Llama 3-70B 83.5%
Qwen2.5 72B Instruct 82.3%
GPT-3.5-turbo 81.6%
phi-3-medium 14B 81.5%
Phi 3 Small 8k Instruct 81.5%
Qwen2.5-Coder-32B 80.8%
Llama 2-70B 80.2%
75808590

Sponsored placements are available on pages like this one. Advertise on Noometry

All results

WinoGrande results by model
#	Model	Provider	Score	Setting	Source
1	Llama 3.1-405B	Meta	89.2%		Epoch AI
2	Claude 3 Opus	Anthropic	88.5%		Epoch AI
3	GPT-4	OpenAI	87.5%		Epoch AI
4	GPT-4	OpenAI	87.5%		Epoch AI
5	Falcon-180B	Technology Innovation Institute	87.1%		Epoch AI
6	DeepSeek-V2 (MoE-236B, May 2024)	DeepSeek	86.3%		Epoch AI
7	DeepSeek-V3	DeepSeek	85.2%		Epoch AI
8	Deepseek Coder v2	DeepSeek	83.7%		Epoch AI
9	Llama 3-70B	Meta	83.5%		Epoch AI
10	Qwen2.5 72B Instruct	Alibaba (Qwen)	82.3%		Epoch AI
11	GPT-3.5-turbo	OpenAI	81.6%		Epoch AI
12	phi-3-medium 14B	Microsoft	81.5%		Epoch AI
13	Phi 3 Small 8k Instruct	Microsoft	81.5%		Epoch AI
14	Qwen2.5-Coder-32B	Alibaba (Qwen)	80.8%		Epoch AI
15	Llama 2-70B	Meta	80.2%		Epoch AI
16	Gemma 7B	Google	79%		Epoch AI
17	Falcon 2 11B	Technology Innovation Institute	78.3%		Epoch AI
18	Nemotron-4 15B	NVIDIA	78%		Epoch AI
19	Mixtral 8x7B	Mistral AI	77.2%		Epoch AI
20	Falcon-40B	Technology Innovation Institute	76.9%		Epoch AI
21	Llama 2-34B	Meta	76.7%		Epoch AI
22	Llama 3-8B	Meta	75.7%		Epoch AI
23	Mistral 7B	Mistral AI	75.3%		Epoch AI
24	Claude 3 Sonnet	Anthropic	75.1%		Epoch AI
25	Claude 3 Haiku	Anthropic	74.2%		Epoch AI
26	Phi-1.5	Microsoft	73.4%	5	Epoch AI
27	Llama 13b	Meta	73%		Epoch AI
28	Qwen2.5-Coder (1.5B)	Alibaba (Qwen)	72.9%		Epoch AI
29	Llama 2-13B	Meta	72.8%		Epoch AI
30	Yi 6B	01.AI	71.3%		Epoch AI
31	Phi 3 Mini 4k Instruct	Microsoft	70.8%		Epoch AI
32	Llama 2-7B	Meta	69.2%		Epoch AI
33	Falcon-7B	Technology Innovation Institute	67.2%		Epoch AI
34	INTELLECT-1	Hugging Face	65.8%		Epoch AI
35	Gemma 2B	Google	65.4%		Epoch AI
36	StarCoder 2 15B	NVIDIA	64.3%		Epoch AI
37	DeepSeek Coder 33B	DeepSeek	62%		Epoch AI
38	Dolly 2.0-12b	Databricks	61.8%		Epoch AI
39	DeepSeek Coder 6.7B	DeepSeek	57.6%		Epoch AI
40	StarCoder 2 3B	NVIDIA	57.1%		Epoch AI
41	StarCoder 2 7B	NVIDIA	57.1%		Epoch AI
42	Phi-2	Microsoft	54.7%		Epoch AI
43	DeepSeek Coder 1.3B	DeepSeek	53.3%		Epoch AI

Compare the leaders

Other reasoning benchmarks

Frequently asked questions

What does WinoGrande measure?

Pronoun-resolution problems that need commonsense reasoning.

Which model has the highest WinoGrande score?

As of October 2026, Llama 3.1-405B has the highest published WinoGrande score on Noometry at 89.2%, out of 43 models with results.

What is the best open-weight model on WinoGrande?

Llama 3.1-405B has the highest WinoGrande accuracy among open-weight models at 89.2%, ranking 1 of 43 overall.