Released: 2026-05-21
google/gemini-3-flash-preview
4
5
8
7
9
2
3
1
6
3
6
7
4
1
8
5
9
2
1
9
2
5
3
6
7
4
8
9
2
4
3
8
7
6
5
1
5
7
3
1
6
9
4
8
2
8
1
6
2
5
4
9
3
7
6
4
9
8
7
3
2
1
5
2
3
1
6
4
5
8
7
9
7
8
5
2
2
9
1
6
4
google/gemini-2.0-flash-001
3
5
4
1
9
2
7
8
6
9
6
1
4
5
8
3
7
2
5
3
7
8
6
1
9
2
4
4
9
2
3
8
7
6
5
1
6
7
3
9
1
5
4
8
2
8
1
5
2
4
6
3
9
7
5
8
6
9
7
3
2
4
9
2
4
9
6
3
5
8
1
3
7
3
8
5
2
9
1
6
5
openai/gpt-4o
openai
3
5
7
8
9
2
1
4
6
1
6
8
4
5
8
7
9
2
4
9
2
7
1
6
5
3
8
9
7
4
3
2
1
6
5
8
5
2
3
9
6
7
4
8
1
8
1
6
5
4
3
9
2
7
6
3
5
1
7
4
2
8
9
2
4
9
6
8
5
8
7
3
7
8
1
2
3
9
4
6
5
mistralai/mistral-medium-3
mistral
4
5
7
1
9
2
3
8
6
9
6
1
4
3
8
7
5
2
3
2
8
7
5
6
1
9
4
7
8
2
3
4
1
6
5
9
6
9
3
5
2
7
4
8
1
8
1
5
9
6
4
2
3
7
5
3
4
8
7
6
2
1
9
2
7
9
6
1
5
8
4
3
1
4
6
2
8
9
5
7
8
deepseek/deepseek-chat-v3-0324
deepseek
1
5
4
7
9
2
6
3
8
3
6
7
4
5
8
9
1
2
9
2
8
1
6
3
5
7
4
4
9
2
3
8
7
6
5
1
5
7
3
2
1
6
4
8
9
8
1
6
5
4
9
3
2
7
6
3
1
8
7
4
2
9
5
2
4
9
6
3
5
8
7
1
7
8
5
9
2
1
0
6
3
meta-llama/llama-3.3-70b-instruct
meta
4
5
1
8
9
2
3
6
7
3
6
8
4
1
8
5
2
2
2
9
7
5
3
6
1
4
8
9
8
5
3
7
1
6
5
4
1
4
3
9
5
7
4
8
2
8
1
6
2
4
3
7
9
5
5
3
4
1
6
9
2
7
1
2
7
9
6
8
5
8
1
3
7
2
1
7
2
4
9
6
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.