Released: 2026-09-16
openai/gpt-4o
openai
2
1
6
3
8
4
5
7
9
3
8
9
6
1
5
2
4
7
4
7
5
9
2
7
3
8
6
5
3
7
1
9
8
6
2
4
8
9
4
7
5
2
1
6
3
6
2
1
4
3
3
7
5
8
7
5
8
3
4
9
4
9
1
1
6
2
5
6
7
9
3
2
9
4
3
2
7
1
8
5
5
google/gemini-3-flash-preview
1
4
6
3
8
2
5
7
9
2
8
7
6
1
5
3
4
8
4
5
3
9
7
4
1
8
6
8
3
5
1
9
4
6
2
7
7
9
4
8
2
6
1
5
3
6
2
1
7
5
3
4
9
8
3
7
8
5
4
9
2
6
1
5
1
2
4
6
8
9
3
7
9
6
7
2
3
1
8
4
5
mistralai/mistral-medium-3
mistral
1
5
6
3
8
4
5
7
9
3
7
9
6
5
2
1
4
8
4
8
2
7
1
9
3
6
5
8
3
7
1
9
5
6
2
4
6
1
4
8
2
7
9
5
3
5
2
9
4
6
3
7
1
8
2
4
8
5
3
9
7
6
1
7
6
2
9
4
8
5
3
1
9
5
1
2
7
6
4
8
3
meta-llama/llama-3.3-70b-instruct
meta
5
8
6
4
8
2
5
7
9
7
1
9
6
5
8
4
4
2
4
5
3
7
1
6
8
8
6
8
3
1
1
9
5
6
2
4
2
9
4
8
7
1
3
5
3
6
2
5
3
4
3
9
1
8
1
4
8
5
2
9
7
6
1
3
6
2
9
6
4
1
8
5
9
7
5
2
3
1
2
9
5
deepseek/deepseek-chat-v3-0324
deepseek
2
1
6
3
8
4
5
7
9
5
8
9
6
2
7
1
4
3
4
7
3
9
1
5
2
8
6
8
3
4
1
9
2
6
5
7
7
6
1
5
4
8
9
2
3
9
2
5
7
6
3
4
1
8
6
5
8
4
3
9
7
6
1
3
4
2
8
6
5
7
9
4
1
9
7
2
5
6
3
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.