Released: 2026-06-07
openai/gpt-4o
openai
3
7
6
5
2
1
8
9
4
8
4
2
6
9
3
5
7
1
5
1
9
8
7
4
6
2
3
7
5
3
9
8
2
4
6
1
2
6
8
3
4
5
1
3
7
9
9
4
6
1
7
3
5
2
6
8
1
4
5
2
7
1
9
4
2
5
7
6
8
9
1
8
1
3
7
2
3
9
2
4
5
google/gemini-3-flash-preview
3
7
6
5
2
1
8
9
4
4
1
2
8
9
3
5
7
6
8
9
5
4
7
6
2
3
1
7
5
1
2
8
9
4
6
3
2
6
8
3
4
5
9
1
7
9
3
4
6
1
7
8
5
2
1
8
3
9
5
2
7
4
6
5
2
9
7
6
8
3
1
4
6
4
7
1
3
4
1
2
5
mistralai/mistral-medium-3
mistral
3
7
6
8
1
5
9
2
4
4
1
2
4
9
3
5
8
6
9
4
5
2
6
7
1
3
8
7
5
9
3
8
4
2
6
1
2
6
8
5
4
1
3
9
7
1
3
4
6
7
9
8
5
2
6
8
3
9
5
2
4
7
1
5
2
1
7
6
8
3
4
9
8
9
7
1
3
6
5
2
4
meta-llama/llama-3.3-70b-instruct
meta
3
7
6
5
1
9
2
8
4
5
9
1
8
9
3
5
6
7
4
3
5
2
6
8
9
1
5
7
5
3
1
8
4
6
6
2
2
1
8
3
4
5
7
9
7
6
4
4
6
7
2
1
5
2
9
8
2
4
5
2
3
7
1
1
2
9
7
3
8
4
5
6
8
6
7
9
3
1
5
4
3
deepseek/deepseek-chat-v3-0324
deepseek
3
7
6
5
2
1
8
9
4
8
4
2
6
9
3
5
7
1
5
1
9
8
7
4
2
3
6
7
5
1
2
8
9
4
6
3
2
6
8
3
4
5
9
1
7
9
3
4
6
1
7
8
5
2
6
8
3
4
5
2
7
1
9
4
2
5
7
6
8
1
3
9
1
9
7
9
3
6
4
2
8
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.