Released: 2026-05-30
google/gemini-3-flash-preview
9
3
1
7
5
2
8
6
4
8
2
6
1
4
0
0
0
3
4
7
5
6
8
3
9
1
2
6
5
3
8
9
4
2
7
1
7
4
2
5
1
6
3
8
9
1
9
8
2
3
7
4
5
6
2
6
4
3
7
1
5
9
8
3
1
7
5
6
9
2
4
7
5
8
9
4
2
8
1
3
0
openai/gpt-4o
openai
2
3
1
9
7
2
8
6
4
8
4
9
5
6
1
2
7
3
6
7
5
1
8
3
9
4
2
1
2
6
7
5
4
3
9
8
5
4
2
3
1
6
7
8
9
7
9
8
2
3
8
4
5
1
4
6
3
9
2
7
5
1
8
9
1
7
8
6
9
6
4
5
3
8
9
4
9
5
1
3
6
meta-llama/llama-3.3-70b-instruct
meta
5
3
9
8
6
2
1
6
4
8
1
6
5
9
7
4
3
2
4
7
2
1
8
3
6
5
9
3
5
1
4
2
9
8
7
6
2
9
8
7
1
6
5
4
3
6
4
5
3
7
1
2
9
8
9
8
7
2
5
4
3
1
6
1
2
3
6
4
8
7
5
9
7
6
4
9
3
5
1
2
8
google/gemini-2.0-flash-001
9
3
1
5
7
2
8
6
4
8
2
6
1
4
5
9
7
3
4
7
5
9
8
3
6
1
2
3
1
8
6
9
4
2
5
7
5
4
2
7
1
6
3
8
9
6
9
7
2
3
8
4
1
5
7
6
4
3
2
1
5
9
8
1
5
3
8
6
9
7
4
2
2
8
9
4
5
7
1
3
6
mistralai/mistral-medium-3
mistral
5
3
1
9
7
2
8
6
4
8
2
4
5
9
1
6
7
3
6
7
9
1
8
3
4
5
2
3
1
7
8
5
4
9
2
6
4
5
2
3
1
6
7
8
9
7
9
6
2
3
8
5
4
1
9
6
8
7
4
5
5
3
8
2
8
3
1
6
9
2
4
7
1
4
5
4
2
7
1
9
5
deepseek/deepseek-chat-v3-0324
deepseek
1
3
9
5
7
2
8
6
4
8
2
6
1
4
9
7
5
3
5
7
4
6
8
3
9
1
2
6
1
5
7
9
4
2
3
8
3
4
2
9
1
6
5
8
7
7
9
8
2
3
5
6
4
1
9
6
1
3
2
7
5
4
8
2
5
7
8
6
9
3
4
1
4
8
3
4
5
1
1
2
6
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.