Released: 2026-06-09
openai/gpt-4o
openai
6
9
8
4
3
2
5
7
1
4
1
7
5
6
9
2
8
3
5
3
2
8
1
7
6
9
4
8
7
1
2
5
3
4
2
6
9
2
6
9
7
4
1
5
7
1
5
3
6
8
7
9
4
2
7
4
9
3
2
5
8
1
0
3
6
5
7
8
6
7
3
2
2
8
4
1
9
4
3
6
9
google/gemini-3-flash-preview
2
9
6
4
3
8
5
7
1
4
7
3
5
6
1
2
8
9
1
8
5
2
9
7
6
3
4
8
5
1
6
4
3
7
9
2
6
2
4
9
7
5
1
5
3
7
3
9
1
8
2
4
6
5
9
4
8
7
2
5
3
1
6
3
1
7
8
5
6
9
4
2
5
6
2
3
9
4
8
7
1
meta-llama/llama-3.3-70b-instruct
meta
5
9
8
1
3
2
5
7
6
4
1
3
8
6
5
2
8
9
2
6
5
4
8
7
1
3
4
8
5
1
7
9
3
4
2
6
9
2
7
9
7
1
8
5
3
1
8
4
5
2
6
9
6
7
6
4
2
3
1
8
7
9
5
3
7
9
2
5
4
6
1
2
7
3
6
6
4
9
5
8
1
mistralai/mistral-medium-3
mistral
6
9
8
1
4
2
5
7
3
4
7
5
3
6
5
9
8
1
1
3
2
5
8
7
6
3
4
8
5
1
2
3
3
4
1
6
7
2
6
9
7
4
3
5
8
9
4
3
8
5
1
2
6
7
2
4
7
6
2
8
1
9
5
3
8
9
7
1
6
8
4
2
5
6
4
3
9
9
7
2
8
deepseek/deepseek-chat-v3-0324
deepseek
2
9
6
8
4
1
5
7
3
4
7
3
2
6
5
1
8
9
5
1
8
3
7
9
2
6
4
8
5
1
7
2
3
4
9
6
9
2
4
6
8
7
3
5
1
6
3
7
1
5
4
8
2
9
7
4
9
5
2
8
6
3
1
3
8
5
4
1
6
9
2
7
1
6
2
9
3
7
4
8
5
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.