Released: 2026-05-15
google/gemini-3-flash-preview
3
8
9
5
7
1
2
6
4
2
7
5
9
6
3
8
4
1
1
6
4
8
4
2
5
7
3
8
1
2
7
4
5
6
3
9
4
5
7
6
3
9
1
2
8
9
3
6
1
2
4
4
5
7
7
2
1
4
9
8
3
5
6
5
4
3
2
8
6
7
1
9
6
9
8
3
1
7
4
3
2
openai/gpt-4o
openai
5
8
9
3
2
1
6
7
4
2
3
5
9
6
7
8
4
1
1
6
7
8
4
2
5
9
3
3
1
4
6
7
5
2
8
9
7
5
8
1
3
9
4
2
6
9
2
6
5
8
4
1
3
7
4
7
2
7
9
3
3
5
6
8
9
3
4
5
6
7
1
5
6
4
1
2
1
8
9
6
2
deepseek/deepseek-chat-v3-0324
deepseek
4
8
9
3
5
1
2
6
7
2
3
5
9
6
7
4
8
1
1
6
7
8
2
4
5
9
3
3
1
2
7
8
5
6
4
9
7
5
8
6
3
9
1
2
4
9
4
6
2
1
4
3
5
7
5
2
1
4
9
3
7
8
6
8
7
3
5
4
6
9
1
2
6
9
4
1
7
8
8
3
2
mistralai/mistral-medium-3
mistral
7
8
9
2
4
1
6
5
4
2
3
5
9
6
7
4
8
1
1
4
6
8
5
3
5
9
3
3
1
2
7
8
5
9
6
4
5
7
8
1
3
6
2
4
9
4
9
6
3
2
4
8
5
7
9
2
4
5
9
8
1
3
6
8
5
3
6
1
2
7
4
9
6
6
7
4
7
9
3
2
8
x-ai/grok-3-mini-beta
x-ai
4
8
9
7
2
1
6
3
5
2
3
5
9
6
8
7
4
1
7
6
1
8
4
2
5
9
3
9
1
2
4
7
5
3
6
8
8
7
4
6
3
9
1
5
2
1
5
6
3
8
4
2
7
7
3
2
7
1
5
7
8
3
6
5
4
3
2
1
6
9
8
4
6
1
8
5
9
3
4
2
7
google/gemini-2.0-flash-001
3
8
9
4
5
1
7
6
2
2
4
5
9
6
7
8
3
1
7
6
1
8
2
3
5
9
4
9
1
2
7
8
5
6
4
3
4
7
8
6
3
9
1
2
5
5
3
6
1
2
4
9
8
7
8
2
4
5
9
1
3
7
6
5
9
3
2
7
6
4
1
8
6
1
7
3
4
8
2
5
9
meta-llama/llama-3.3-70b-instruct
meta
5
8
9
4
2
1
7
6
3
2
4
5
9
6
8
1
3
1
1
7
3
8
5
6
4
2
3
9
1
4
7
8
5
2
8
6
8
5
2
3
3
9
6
4
1
6
3
1
2
4
7
5
9
8
4
9
8
1
3
2
6
5
7
3
2
6
5
1
4
8
7
9
7
6
7
6
9
3
1
5
2
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.