List your company
ResearchFormerly WizWand

General Reasoning Average on PostTrainBench

44.81Average (%)Official Instruct Model
1.878813.024424.1735.3156May 31, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Official Instruct ModelProtocol=Official Instruct, Backbone=SmolLM3-3B2026.0544.81
GLM-4.7 & ANDES (Ours)Protocol=Proposed Method, Backbone=SmolLM3-3B2026.0531.91
Opus-4.6 (1M)Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.0528.52
Opus-4.7 (xHigh)Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.0525.28
Opus-4.6Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.0523.16
Gemini-3.1-ProProtocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.0522.08
GPT-5.2Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.0521.26
GPT-5.4 (High)Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.0520.72
GLM-4.7 (Scaffold-only)Protocol=Proposed Method, Backbone=SmolLM3-3B2026.0520.12
MiniMax-M2.5Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.059.04
GPT-5.1-Codex-MaxProtocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.058.21
MiniMax-M2.1Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.055.3
Base Model (SmolLM3-3B)Protocol=Zero-Shot, Backbone=SmolLM3-3B2026.054.52
Kimi-K2-ThinkingProtocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.054.52
Qwen3-MaxProtocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.054.5
Sonnet-4.5Protocol=Agent-Post-Trained, Backbone=SmolLM3-3B2026.053.99
GLM-4.7 (OpenCode)Protocol=Proposed Method, Backbone=SmolLM3-3B2026.053.53