Released: 2026-06-13
openai/gpt-4o
openai
5
4
3
9
6
8
1
7
2
7
2
6
3
5
4
9
8
1
8
9
1
2
1
7
6
5
4
4
3
7
8
2
6
5
9
1
6
8
2
7
9
5
3
1
4
1
5
9
4
3
9
7
2
8
9
1
4
6
3
2
8
4
7
3
6
5
1
8
9
7
2
9
6
7
8
5
4
1
2
3
6
google/gemini-3-flash-preview
3
4
8
9
6
5
1
7
2
1
2
6
3
7
8
9
4
5
7
9
1
4
1
2
8
5
6
4
3
9
8
5
1
2
6
7
8
5
2
7
4
6
3
1
9
1
6
7
3
2
9
5
4
8
5
8
4
6
3
7
3
9
1
2
1
5
1
8
4
7
3
6
6
7
3
5
9
2
8
2
4
mistralai/mistral-medium-3
mistral
8
3
7
4
6
9
1
7
2
1
2
6
8
7
3
9
4
5
5
9
4
2
1
8
6
5
3
4
3
9
7
5
2
8
6
1
7
8
2
9
4
6
3
1
9
9
6
3
1
2
5
4
3
8
2
1
4
6
3
7
5
8
9
3
4
5
3
8
1
7
2
6
6
7
8
5
9
4
2
1
3
meta-llama/llama-3.3-70b-instruct
meta
5
1
9
7
6
4
1
7
3
2
8
6
1
5
3
9
4
7
3
4
7
9
1
2
5
8
6
1
5
3
4
9
7
2
6
8
9
6
2
8
3
1
7
5
4
7
3
8
5
2
6
4
9
1
8
9
1
6
7
5
3
2
4
4
7
5
3
8
9
6
1
2
6
2
4
1
4
8
7
3
9
deepseek/deepseek-chat-v3-0324
deepseek
5
4
3
8
6
9
1
7
2
7
2
6
3
5
4
9
8
1
8
9
1
2
1
7
4
5
3
4
3
7
9
2
8
5
6
1
9
8
2
7
4
5
3
1
6
1
6
5
1
9
3
2
4
8
2
1
4
6
3
2
8
9
7
3
5
8
4
8
1
7
2
9
6
7
9
5
7
6
1
3
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.