Released: 2026-07-12
openai/gpt-4o
openai
2
9
1
7
4
8
6
5
3
6
7
8
5
1
3
9
4
2
5
4
3
9
2
6
1
7
8
1
6
4
2
3
5
7
9
8
8
5
7
4
6
9
2
3
1
9
3
2
1
7
8
5
6
4
3
2
6
8
9
1
4
7
5
7
1
5
6
5
4
8
2
9
4
8
9
3
5
7
3
1
6
google/gemini-3-flash-preview
2
9
1
4
7
8
6
5
3
6
7
5
2
1
3
9
4
8
3
4
8
9
5
6
2
7
1
1
6
4
2
3
5
7
9
8
8
5
7
4
6
9
1
3
2
9
2
3
1
8
7
5
6
4
4
1
6
3
8
2
7
9
5
7
3
9
6
4
1
8
2
5
5
8
2
7
9
4
3
1
6
mistralai/mistral-medium-3
mistral
2
9
1
4
7
8
6
5
3
8
7
5
5
2
3
9
4
1
3
4
6
9
1
5
2
8
7
7
6
8
2
3
1
4
9
5
9
5
7
8
6
4
1
3
2
4
3
2
1
9
7
5
6
8
1
2
6
3
8
9
7
4
5
5
8
9
6
4
2
8
2
3
6
1
3
5
7
8
2
1
4
meta-llama/llama-3.3-70b-instruct
meta
2
9
1
8
4
5
6
5
7
8
3
5
9
1
3
9
6
2
4
1
6
3
2
6
5
8
9
6
5
2
4
3
1
3
9
8
3
8
9
5
6
2
1
4
5
1
4
3
1
9
8
5
2
6
5
2
8
6
5
4
7
1
3
9
6
4
7
8
9
2
5
1
7
1
3
2
5
7
4
6
8
deepseek/deepseek-chat-v3-0324
deepseek
2
9
1
8
4
7
6
5
3
6
7
5
9
2
3
1
8
4
3
4
8
5
1
6
2
7
9
1
6
4
2
3
5
7
9
8
8
5
7
4
6
9
3
2
1
9
2
3
1
7
8
5
4
6
7
1
6
3
8
2
4
6
5
4
3
9
6
5
1
8
2
7
5
8
2
7
9
4
3
1
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.