Released: 2026-08-21
openai/gpt-4o
openai
7
5
1
8
2
4
9
3
6
3
9
6
5
1
7
8
2
4
8
2
4
9
6
5
1
7
3
4
6
5
7
3
9
2
8
1
1
3
2
4
5
8
6
7
9
9
8
7
1
4
6
5
5
2
2
4
3
6
8
1
7
9
5
5
7
8
3
9
2
4
6
5
6
1
9
2
7
3
3
1
8
google/gemini-3-flash-preview
7
1
5
4
2
6
8
9
3
2
9
6
3
1
8
5
7
4
8
4
3
9
7
5
1
6
2
5
6
4
7
3
9
2
8
1
1
3
2
8
5
4
9
7
6
9
8
7
1
6
2
4
5
3
3
4
9
6
8
1
7
2
5
6
7
1
2
9
3
5
4
8
4
5
8
5
4
7
3
2
9
mistralai/mistral-medium-3
mistral
7
5
9
2
3
6
8
4
1
3
2
6
8
1
9
7
5
4
8
1
4
3
7
5
6
9
2
5
6
8
7
4
9
2
3
1
1
3
2
4
5
8
9
7
6
9
8
7
1
2
3
4
5
6
2
4
3
6
8
1
5
7
9
6
7
1
5
9
2
3
8
4
4
9
5
3
6
7
1
2
8
deepseek/deepseek-chat-v3-0324
deepseek
7
1
2
4
3
8
5
6
9
3
5
6
2
1
9
7
8
4
8
9
4
7
6
5
1
2
3
4
6
5
7
8
9
2
3
1
1
3
9
6
5
2
4
7
8
2
8
7
1
4
3
6
5
9
5
4
3
6
8
2
9
1
7
6
7
8
3
9
1
2
4
5
9
2
1
5
7
4
3
8
6
meta-llama/llama-3.3-70b-instruct
meta
7
2
5
3
6
8
1
4
9
1
9
6
4
1
3
5
8
4
8
5
3
2
4
5
6
9
1
4
1
9
7
3
9
2
6
1
5
3
1
8
5
2
4
7
6
6
8
4
1
9
7
3
5
2
3
4
2
6
8
1
7
1
5
9
6
7
5
2
4
8
3
1
2
1
8
9
7
6
3
2
4
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.