Released: 2026-07-18
openai/gpt-4o
openai
6
9
5
2
3
7
8
4
1
3
4
2
5
6
9
7
1
8
7
8
1
4
2
8
6
5
9
4
6
8
3
5
1
9
2
7
5
1
3
7
9
2
4
8
6
9
7
2
8
4
6
5
3
1
2
9
6
1
1
5
3
7
4
8
3
7
9
5
4
2
6
5
1
5
4
6
7
3
1
9
2
google/gemini-3-flash-preview
6
9
1
2
5
8
7
4
3
3
2
5
9
4
6
1
7
8
7
4
8
1
3
7
6
5
9
4
6
2
3
8
5
9
1
7
5
1
3
7
9
2
8
6
4
8
7
9
4
6
1
5
3
2
2
8
6
5
1
3
4
9
7
7
3
4
8
2
4
4
1
6
1
5
4
6
7
9
3
8
2
mistralai/mistral-medium-3
mistral
6
8
5
2
3
9
7
4
1
3
9
4
7
6
5
1
2
8
7
2
1
1
4
8
6
5
9
4
6
9
3
5
1
8
9
7
5
1
3
8
9
7
4
6
2
8
7
2
4
6
3
5
1
9
2
4
7
5
1
6
3
8
5
9
3
6
9
2
4
7
3
6
1
5
8
6
7
3
2
9
4
meta-llama/llama-3.3-70b-instruct
meta
6
8
5
2
4
3
1
4
9
3
4
2
8
5
1
6
9
7
1
9
7
4
6
5
8
3
2
4
3
6
5
8
2
9
1
7
8
1
4
3
9
7
5
6
2
5
2
9
1
7
6
4
8
3
2
6
1
9
3
8
7
5
4
9
5
8
7
2
4
3
1
6
7
4
3
6
1
9
2
5
8
deepseek/deepseek-chat-v3-0324
deepseek
6
9
1
2
5
7
8
4
3
3
2
5
4
6
9
7
1
8
7
4
8
1
3
8
6
5
9
4
6
2
3
8
5
9
2
7
5
1
3
7
9
2
4
8
6
8
7
9
6
4
1
5
3
2
2
8
6
9
1
3
4
7
5
9
3
7
5
2
4
1
6
8
1
5
4
6
7
8
2
9
3
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.