Released: 2026-06-15
openai/gpt-4o
openai
4
3
2
5
6
7
9
1
8
7
9
6
8
1
4
2
5
3
8
5
1
9
2
3
4
7
6
6
8
9
3
7
1
5
4
2
1
7
5
4
8
6
3
2
9
3
2
4
2
5
9
6
8
7
5
1
7
6
4
8
7
3
9
2
4
8
7
3
9
1
6
5
9
6
3
1
2
5
8
7
4
mistralai/mistral-medium-3
mistral
3
4
2
7
6
9
8
1
5
7
5
6
8
1
3
2
9
4
8
9
1
9
2
4
5
7
3
4
6
9
3
7
8
1
5
2
1
3
8
4
8
7
3
2
9
5
2
7
2
5
1
6
4
8
6
8
7
1
3
2
9
4
5
2
1
4
9
4
3
1
6
5
9
7
3
5
9
6
4
8
7
deepseek/deepseek-chat-v3-0324
deepseek
4
9
2
5
6
7
3
1
8
5
7
6
8
1
3
2
9
4
8
3
1
9
2
4
5
7
6
6
2
9
3
4
8
7
5
1
1
5
7
6
8
2
4
2
9
3
4
8
2
5
1
6
7
9
9
8
7
4
3
5
8
6
2
2
1
4
7
9
6
1
6
5
7
6
3
1
2
9
4
8
7
meta-llama/llama-3.3-70b-instruct
meta
5
3
2
1
6
8
4
9
7
6
8
6
8
1
5
2
3
4
8
9
1
9
3
4
5
7
2
4
2
9
3
7
1
8
5
6
1
5
8
4
8
2
7
2
9
7
1
4
2
5
9
6
8
3
9
6
7
5
4
3
1
2
8
2
4
3
6
9
7
5
1
5
3
7
5
8
2
6
9
4
1
google/gemini-3-flash-preview
7
1
2
5
6
3
8
9
4
9
4
6
8
1
7
2
5
3
8
3
5
9
2
4
1
7
6
5
2
9
3
7
6
4
1
8
1
6
3
4
8
9
7
2
5
4
7
8
2
5
1
6
3
9
3
5
7
6
9
8
5
4
2
2
8
4
7
3
5
1
6
9
6
9
1
1
4
2
3
8
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.