Released: 2026-06-04
openai/gpt-4o
openai
7
6
9
3
5
2
1
8
4
2
4
8
9
7
1
6
5
3
5
1
3
8
6
4
7
9
2
3
9
2
1
8
5
9
4
6
6
5
4
7
2
9
3
2
7
1
7
5
6
4
3
2
8
9
4
3
7
2
9
6
5
1
9
8
2
6
5
1
7
4
3
1
9
8
1
4
3
8
6
7
5
google/gemini-3-flash-preview
3
6
9
8
5
2
1
7
4
4
2
8
9
7
1
6
5
3
5
7
1
4
6
3
8
9
2
7
9
3
5
8
4
2
4
6
6
8
2
7
1
9
3
4
5
1
5
4
3
2
6
9
8
7
8
3
7
2
4
5
5
1
9
9
4
5
1
8
7
7
3
2
2
1
6
6
9
8
4
6
1
meta-llama/llama-3.3-70b-instruct
meta
5
1
9
7
2
8
6
3
4
2
8
6
3
5
1
7
9
4
4
3
7
9
6
5
2
1
8
8
5
2
1
4
3
9
7
6
6
9
1
8
7
2
5
4
3
3
7
4
6
9
5
1
8
2
1
6
8
5
3
7
4
2
9
9
4
5
2
1
6
8
3
7
7
2
3
4
8
9
6
5
1
mistralai/mistral-medium-3
mistral
3
7
6
9
5
2
1
8
4
4
2
8
6
7
1
9
5
3
5
1
9
8
6
3
7
2
4
2
9
3
5
8
7
4
1
6
6
4
5
7
1
9
3
2
8
1
8
7
2
3
4
5
6
9
9
3
7
2
4
6
8
1
5
8
5
1
3
2
8
6
3
7
7
6
2
1
9
5
3
4
8
deepseek/deepseek-chat-v3-0324
deepseek
3
7
6
9
5
2
1
8
4
9
2
8
4
7
1
6
5
3
5
1
4
8
6
3
7
9
2
2
9
5
3
8
7
4
6
1
6
4
3
7
1
9
2
8
5
1
8
7
5
2
6
9
3
4
4
3
7
2
6
8
5
1
9
8
6
2
1
4
5
3
7
8
7
5
1
6
9
4
8
2
6
Each model receives the same Sudoku puzzle with a partially filled grid. Models are tasked with solving the puzzle by filling in all empty cells following Sudoku rules, and their solutions are evaluated without feedback or retries.