Released: 2026-05-24
google/gemini-3-flash-preview
9
7
1
5
6
8
2
3
4
6
2
4
7
9
3
5
8
1
8
3
5
4
2
1
6
9
7
1
9
6
8
5
4
7
2
3
2
4
7
6
3
7
8
9
5
3
8
5
9
7
2
4
1
6
4
1
2
3
5
6
9
7
8
5
6
3
1
8
7
4
2
9
7
5
8
2
4
9
1
6
3
openai/gpt-4o
openai
9
3
1
5
7
6
8
4
2
6
2
7
7
4
1
5
3
9
8
4
5
3
2
9
6
1
7
1
7
6
8
5
4
2
3
9
2
5
4
6
3
7
1
9
8
3
8
9
2
1
5
7
1
6
4
9
2
9
6
3
3
7
8
5
6
8
1
8
2
4
5
3
7
1
3
4
9
8
9
2
5
google/gemini-2.0-flash-001
9
3
4
5
1
6
2
8
7
6
2
1
7
9
8
5
4
3
8
7
5
4
2
3
6
9
1
3
9
6
8
7
4
1
5
2
2
1
7
6
3
5
8
9
4
5
8
4
2
9
1
7
3
6
4
6
2
3
5
9
1
7
8
1
5
9
4
8
7
3
6
2
7
9
3
1
6
2
4
5
5
mistralai/mistral-medium-3
mistral
9
3
4
5
1
6
8
7
2
6
2
8
7
4
9
5
3
1
1
7
5
8
2
3
6
4
9
3
5
6
9
7
4
1
8
2
2
4
1
6
3
8
7
9
5
7
8
9
2
5
1
3
6
4
4
6
2
3
9
7
8
5
1
5
9
7
1
8
2
4
6
3
8
1
3
4
6
5
9
2
7
deepseek/deepseek-chat-v3-0324
deepseek
9
7
8
5
6
1
2
3
4
6
2
1
7
4
3
5
8
9
3
4
5
9
2
8
6
1
7
1
9
6
8
5
4
7
2
3
2
5
7
6
3
2
1
9
8
5
8
3
2
7
9
4
1
6
4
6
2
3
9
5
8
7
1
8
3
9
1
8
7
4
6
2
7
1
4
4
1
6
9
2
5
meta-llama/llama-3.3-70b-instruct
meta
9
5
7
5
8
6
2
3
1
6
2
4
7
9
3
5
8
5
8
3
5
2
1
4
6
9
7
1
4
6
3
5
4
8
2
9
2
9
3
8
3
1
7
9
5
5
8
1
9
2
7
4
1
3
4
1
9
6
7
8
3
5
8
3
6
8
1
8
5
9
7
2
7
5
2
4
6
9
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.