Released: 2026-05-25
openai/gpt-4o
openai
1
6
5
2
3
9
7
4
8
2
8
3
7
4
6
1
5
9
7
9
4
6
1
5
2
3
0
3
7
8
5
6
4
9
1
2
9
5
2
1
8
7
6
0
4
6
2
1
9
4
8
3
9
5
5
3
6
8
3
2
4
7
1
4
1
7
3
5
9
8
2
6
8
4
9
0
2
3
5
6
7
google/gemini-3-flash-preview
1
4
6
3
9
2
5
7
8
2
5
3
4
8
6
1
7
9
7
9
8
7
1
5
2
3
6
9
7
1
5
6
3
4
8
2
3
5
4
1
9
8
6
7
2
6
2
8
7
4
1
3
9
5
4
6
2
8
7
9
1
5
3
7
3
5
6
1
4
9
2
8
8
1
9
5
2
3
4
6
7
google/gemini-2.0-flash-001
5
1
6
4
3
7
2
9
8
2
8
3
9
5
6
1
7
4
4
9
7
2
1
5
6
8
3
9
7
8
3
6
4
5
1
2
3
5
4
1
9
2
7
8
6
6
2
1
5
7
8
3
4
9
7
6
2
8
4
9
9
3
1
8
3
5
7
6
1
4
2
9
1
4
9
6
2
3
8
5
7
mistralai/mistral-medium-3
mistral
4
6
7
9
3
2
5
1
8
2
8
3
7
5
6
1
4
9
5
9
1
4
1
5
8
7
6
9
7
8
3
6
4
5
2
1
3
5
2
1
9
7
6
8
4
6
2
4
5
8
9
3
7
1
1
3
6
8
4
9
2
5
7
7
4
8
6
5
1
9
2
3
8
1
9
2
7
3
4
6
5
meta-llama/llama-3.3-70b-instruct
meta
5
1
4
3
7
9
6
8
2
2
8
3
5
4
6
1
7
9
9
6
7
2
1
8
5
3
4
1
7
5
9
6
4
2
8
3
8
5
2
1
3
7
4
9
6
6
3
9
8
5
2
7
1
4
4
2
1
7
9
3
8
6
5
3
9
8
4
2
1
6
5
7
7
4
6
5
8
9
3
2
1
deepseek/deepseek-chat-v3-0324
deepseek
1
6
5
2
3
7
4
9
8
2
8
3
9
4
6
1
7
5
7
9
4
8
1
5
2
3
6
9
7
8
3
6
4
5
1
2
4
5
2
1
9
8
6
7
3
6
2
1
5
7
9
3
9
4
3
4
6
8
5
2
7
8
9
5
3
7
6
8
1
9
2
4
8
1
9
4
2
3
8
6
7
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.