List your company

Code Reasoning on HumanE

84.9AccuracyDenser
69.61273.58177.5581.519Dec 17, 2025

Evaluation results

MethodLinks
DenserModel=GPT-4o2025.1284.9—
Process SupervisionModel=GPT-4o2025.1284.6—
Reflection-CoTModel=GPT-4o2025.1284—
Self-VerificationModel=GPT-4o2025.1283.7—
Tree-of-ThoughtModel=GPT-4o2025.1283.2—
DenserModel=DeepSeek-V32025.1283.1—
Process SupervisionModel=DeepSeek-V32025.1282.9—
Self-ConsistencyModel=GPT-4o2025.1282.5—
Reflection-CoTModel=DeepSeek-V32025.1282.3—
Self-VerificationModel=DeepSeek-V32025.1281.9—
Think-to-ThinkModel=GPT-4o2025.1281.8—
Tree-of-ThoughtModel=DeepSeek-V32025.1281.6—
Self-ConsistencyModel=DeepSeek-V32025.1280.9—
DenserModel=o1-mini2025.1280.8—
Process SupervisionModel=o1-mini2025.1280.7—
Think-to-ThinkModel=DeepSeek-V32025.1280.2—
Reflection-CoTModel=o1-mini2025.1280.1—
Chain-of-ThoughtModel=GPT-4o2025.1279.8—
Self-VerificationModel=o1-mini2025.1279.5—
Tree-of-ThoughtModel=o1-mini2025.1278.9—
Self-ConsistencyModel=o1-mini2025.1278.2—
Chain-of-ThoughtModel=DeepSeek-V32025.1278.2—
DenserBackbone=Qwen3-14B2025.1278.2-58.7
Think-to-ThinkModel=o1-mini2025.1277.5—
Self-VerificationBackbone=Qwen3-14B2025.1276.9142.8
Process SupervisionBackbone=Qwen3-14B2025.1276.3156.3
Chain-of-ThoughtModel=o1-mini2025.1275.8—
TokenSkip (0.9 ratio)Backbone=Qwen3-14B2025.1275.8-53.5
Reflection-CoTBackbone=Qwen3-14B2025.1275.4127.6
Tree-of-ThoughtBackbone=Qwen3-14B2025.1275345.7
Self-ConsistencyBackbone=Qwen3-14B2025.1274.1289.4
Think-to-ThinkBackbone=Qwen3-14B2025.1273.8187.2
Chain-of-Thought (CoT)Backbone=Qwen3-14B2025.1272.90
BeConciseBackbone=Qwen3-14B2025.1271.5-54.2
AbbreWordsBackbone=Qwen3-14B2025.1270.2-50.8