Released: 2026-08-13
openai/gpt-4o
openai
5
3
1
7
4
6
2
9
8
7
6
9
1
3
4
8
5
2
2
4
8
5
9
1
7
1
6
6
8
5
3
2
9
4
7
1
4
7
2
9
5
8
6
3
4
3
1
9
4
1
7
5
8
9
9
5
3
6
7
2
1
6
4
8
2
6
4
1
5
3
7
9
1
9
4
8
6
3
9
2
1
google/gemini-3-flash-preview
5
3
4
7
2
6
1
8
9
1
6
7
9
8
4
3
5
2
8
2
9
5
1
3
7
4
6
6
8
2
3
4
9
5
1
7
4
7
1
8
5
2
6
3
9
3
9
5
1
6
7
4
2
8
9
1
3
6
7
5
2
8
4
2
4
6
4
3
1
8
9
5
7
5
8
8
9
2
9
6
1
mistralai/mistral-medium-3
mistral
5
3
4
7
6
8
1
9
2
1
6
7
3
9
4
8
5
2
8
2
9
1
5
3
7
4
6
6
8
2
4
1
9
3
7
5
4
7
1
6
5
2
8
3
9
3
9
5
8
7
3
2
6
4
9
1
3
6
2
7
5
8
4
2
4
6
9
3
5
1
7
8
7
5
8
8
4
6
9
2
1
deepseek/deepseek-chat-v3-0324
deepseek
5
3
2
7
6
8
1
4
9
7
6
8
9
1
4
3
5
2
1
9
4
5
2
3
7
8
6
6
8
5
1
3
9
4
2
7
4
7
9
6
5
2
8
3
1
2
1
3
4
8
7
6
9
5
9
2
3
6
7
5
2
8
4
3
4
6
2
9
1
5
7
8
8
5
7
3
4
6
9
2
1
meta-llama/llama-3.3-70b-instruct
meta
5
3
4
7
6
8
9
1
2
6
7
2
1
9
4
3
5
8
1
9
8
3
4
2
7
6
5
6
8
3
5
1
9
4
2
7
4
7
5
2
8
6
1
3
9
2
1
9
4
7
3
8
5
6
9
5
1
6
3
7
2
8
4
8
4
6
9
5
1
7
3
2
3
2
7
8
4
5
6
9
1
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.