Share your thoughts, 1 month free Claude Pro on usSee more

Arithmetic Reasoning on GSM8K

97.1Accuracy

DUP

Updated 4d ago

Evaluation Results

Method
DUP 2024.04	97.1	-	-	-	-	-	-
Llama-3.1-405B 2025.12	95.9	-	-	-	-	-	-
Automatic Model Selection with LLMs 2023.05	95.6	-	-	-	-	-	-
Claude 3.5 Sonnet 2025.12	95.6	-	-	-	-	-	-
GHS-TDA 2026.02	95.2	-	-	-	-	-	-
GHS-TDA 2026.02	95.2	-	-	-	-	-	-
AoT 2026.02	95.1	-	-	-	-	-	-
AoT 2026.02	95	-	-	-	-	-	-
ToT 2026.02	95	-	-	-	-	-	-
ToT 2026.02	94.9	-	-	-	-	-	-
GHS-TDA 2026.02	94.8	-	-	-	-	-	-
AoT 2026.02	94.7	-	-	-	-	-	-
CoT 2023.05	94.6	-	-	-	-	-	-
Zero-shot CoT 2024.04	94.6	-	-	-	-	-	-
GPT-4o 2025.12	94.6	-	-	-	-	-	-
GoT 2026.02	94.5	-	-	-	-	-	-
GoT 2026.02	94.5	-	-	-	-	-	-
Zero-shot PS+ 2024.04	94.3	-	-	-	-	-	-
ToT 2026.02	94.2	-	-	-	-	-	-
FoT 2026.02	94.2	-	-	-	-	-	-
PAL 2023.05	94	-	-	-	-	-	-
FoT 2026.02	94	-	-	-	-	-	-
FoT 2026.02	93.9	-	-	-	-	-	-
GoT 2026.02	93.8	-	-	-	-	-	-
AFlow 2026.02	93.6	-	-	-	-	-	-
AFlow 2026.02	93.5	-	-	-	-	-	-
AFlow 2026.02	93.1	-	-	-	-	-	-
LLM-PeerReview-W 2025.12	93	-	-	67.8	-	-	-
LLM-PeerReview 2025.12	92.7	-	-	67.4	-	-	-
GPT-4o 2025.12	92.6	-	-	-	-	-	-
Least-to-Most 2024.04	92.1	-	-	-	-	-	-
CoT-SC 2026.02	92.1	-	-	-	-	-	-
Qwen2.5-7B-Selection 2025.12	92.1	-	-	65.5	-	-	-
CoT-SC 2026.02	92	-	-	-	-	-	-
Self-Refine 2026.02	91.9	-	-	-	-	-	-
GaC 2025.12	91.8	-	-	60.2	-	-	-
Self-Refine 2026.02	91.7	-	-	-	-	-	-
CoT-SC 2026.02	91.5	-	-	-	-	-	-
Qwen2.5-7B-Instruct 2025.12	91.5	-	-	62.7	-	-	-
Smoothie-Global 2025.12	91.5	-	-	60.5	-	-	-
Self-Refine 2026.02	91.4	-	-	-	-	-	-
CoT 2026.02	91.3	-	-	-	-	-	-
CoT 2026.02	90.9	-	-	-	-	-	-
Llama-3.1-8B-Selection 2025.12	90.8	-	-	66.4	-	-	-
Mistral-7B-Selection 2025.12	90.8	-	-	64.7	-	-	-
CoT 2026.02	90.7	-	-	-	-	-	-
GPT-4o 2025.12	90.1	-	-	-	-	-	-
ComplexCoT + SC 2023.06	89.2	-	-	-	-	-	-
Qwen2-7B-Selection 2025.12	88.8	-	-	62.1	-	-	-
Qwen2-7B-Instruct 2025.12	88.5	-	-	56.6	-	-	-
BF16 2026.02	88.48	-	-	-	-	-	-
CoK + SC 2023.06	88.2	-	-	-	-	-	-
OJBKQ 2026.02	87.87	-	-	-	-	-	-
GPT-4o 2025.12	87.7	-	-	-	-	-	-
Analogical Prompting 2026.02	87.6	-	-	-	-	-	-
Manual CoT + SC 2023.06	87.6	-	-	-	-	-	-
Analogical Prompting 2026.02	87.2	-	-	-	-	-	-
Analogical Prompting 2026.02	87	-	-	-	-	-	-
Agent-Forest 2025.12	86.8	-	-	60.1	-	-	-
GPTQ 2026.02	86.8	-	-	-	-	-	-
GPT-4o 2025.12	85.8	-	-	-	-	-	-
Smoothie-Local 2025.12	85.5	-	-	59.8	-	-	-
AWQ 2026.02	85.44	-	-	-	-	-	-
OJBKQ 2026.02	84.29	-	-	-	-	-	-
BF16 2026.02	83.7	-	-	-	-	-	-
CoK 2023.06	83.2	-	-	-	-	-	-
Automatic Model Selection with LLMs 2023.05	82.6	-	-	-	-	-	-
DUP 2024.04	82.3	-	-	-	-	-	-
Manual-CoT 2024.04	81.6	-	-	-	-	-	-
Self-Agreement 2023.11	81.3	-	-	-	-	-	-
Random 2025.12	81.2	-	-	54.2	-	-	-
Theoretical average 2025.12	80.9	-	-	54.1	-	-	-
CoT 2023.05	80.8	-	-	-	-	-	-
Multi-Agents (Debate) 2023.11	80.2	-	-	-	-	-	-
Auto-CoT 2024.04	80.2	-	-	-	-	-	-
AWQ 2026.02	80.15	-	-	-	-	-	-
Zero-shot PS+ 2024.04	79.3	-	-	-	-	-	-
Llama-3.1-8B-Instruct 2025.12	79.3	-	-	53.5	-	-	-
ComplexCoT 2023.06	79.3	-	-	-	-	-	-
PAL 2023.05	79.2	-	-	-	-	-	-
Manual CoT 2023.06	79.1	-	-	-	-	-	-
Zero-shot CoT 2024.04	78.9	-	-	-	-	-	-
Self-consistency 2022.03	78	-	-	-	-	-	-
Least-to-Most 2024.04	77.5	-	-	-	-	-	-
Automatic Model Selection with LLMs 2023.05	74.7	-	-	-	-	-	-
PaLM 540B: Few-Shot-CoT + self consistency 2022.05	74.4	-	-	-	-	-	-
Self-consistency 2022.03	74.4	-	-	-	-	-	-
code-davinci-002 2025.12	72	-	-	-	-	-	-
code-davinci-002 2025.12	71.6	-	-	-	-	-	-
PAL 2023.05	71.5	-	-	-	-	-	-
PaLM 540B: Zero-Shot-CoT + self consistency 2022.05	70.1	-	-	-	-	-	-
Mixed-Few-Shot CoT 2023.11	69	-	-	-	-	-	-
Self-Consistency 2023.11	68	-	-	-	-	-	-
Full FT 2025.06	66.5	-	-	-	-	-	-
DiaBlo 2025.06	66.5	-	-	-	-	-	-
DiaBlo 2025.06	66.3	-	-	-	-	-	-
CoT 2023.05	64.4	-	-	-	-	-	-
QUIP 2026.02	64.4	-	-	-	-	-	-
Mistral-7B-Instruct 2025.12	64.3	-	-	43.5	-	-	-
MiLoRA 2025.06	63.5	-	-	-	-	-	-

Showing 100 of 186 rows