Comparing 2 models

GPT-5.6 Luna vs Hy3

Facts first, then task-specific rankings. A highlighted cell is better only for that row, not universally.

metric
Model 1GPT-5.6 Luna
Model 2Hy3
Best for
Competition Math (#1)
Best Open-Weight LLM for Self-Hosting (#2)
Provider
openai
tencent
Context
1.1MBest
262K
Input / 1M
$0.50
$0.13Best
Output / 1M
$3
$0.53Best
Tool use
yesBest
yesBest
Vision
yesBest
Reasoning
yesBest
yesBest
Open-weight
yesBest
Released
Jul 9, 2026
Jul 6, 2026