Released: 2026-03-26
anthropic/claude-3.7-sonnet
anthropic
9
6
8
5
3
7
4
2
1
1
2
5
9
8
2
6
3
4
3
7
4
4
6
8
5
1
9
2
1
7
6
5
3
9
4
8
6
4
3
8
9
1
7
5
2
7
9
5
3
1
4
2
8
6
4
3
1
7
2
6
8
9
5
5
7
9
2
4
5
3
6
1
8
5
2
1
7
9
4
1
3
google/gemini-2.0-flash-001
9
4
8
5
3
1
7
2
6
1
2
5
9
7
6
8
3
4
3
7
6
4
2
8
5
9
1
8
1
2
6
5
3
4
7
9
6
3
4
2
8
9
1
5
7
7
9
5
8
1
4
2
6
3
4
5
1
7
9
6
3
8
2
2
6
9
3
4
5
1
7
8
5
8
7
1
6
2
9
4
3
google/gemini-3-flash-preview
8
4
7
9
3
5
1
2
6
9
2
5
2
1
7
6
3
4
1
3
6
4
2
8
5
7
9
5
1
4
6
7
3
9
8
2
6
2
3
8
4
9
7
5
1
7
9
8
5
1
2
2
4
3
4
5
1
7
8
6
3
9
2
2
6
9
3
4
1
8
5
7
3
8
2
1
5
4
6
7
9
openai/gpt-4o
openai
8
9
4
5
3
6
1
2
7
1
2
5
9
7
4
6
3
4
3
6
7
4
2
8
5
9
1
2
1
8
6
5
3
7
4
9
4
7
6
2
1
9
3
5
8
7
5
3
8
9
1
2
6
4
4
3
2
7
8
6
9
1
5
6
4
9
3
4
2
8
7
3
5
8
1
1
6
5
4
8
2
mistralai/mistral-medium-3
mistral
1
6
7
9
3
4
8
2
5
2
9
5
3
2
7
6
3
4
3
4
8
4
5
8
9
7
6
4
1
2
6
7
3
5
8
9
5
7
6
8
1
9
3
5
2
7
3
8
5
6
2
2
4
1
4
5
3
7
8
6
1
9
2
6
2
9
3
4
1
7
6
8
8
8
1
1
9
5
4
2
3
deepseek/deepseek-chat-v3-0324
deepseek
6
4
8
5
3
7
9
2
1
1
9
5
2
6
0
6
3
4
2
3
7
4
1
8
5
6
0
5
1
2
6
7
3
4
8
9
3
6
4
8
2
9
7
5
0
7
8
0
0
5
4
2
1
6
4
2
3
7
9
6
8
0
5
0
5
9
3
4
2
1
7
8
0
8
6
1
0
5
3
9
2
meta-llama/llama-3.3-70b-instruct
meta
5
3
4
6
3
9
8
2
1
1
9
5
8
1
6
4
3
4
8
6
2
4
5
8
9
7
6
3
1
7
6
2
3
5
9
8
2
4
9
3
7
9
1
5
6
7
5
1
2
9
4
6
8
3
4
2
8
7
6
5
3
1
9
9
8
6
1
4
7
2
6
5
6
7
3
5
8
1
4
2
9
x-ai/grok-3-mini-beta
x-ai
9
6
1
5
3
7
8
2
4
8
7
5
2
1
4
6
3
9
2
3
4
9
6
8
5
7
1
3
1
8
6
7
5
4
9
2
5
4
2
3
8
9
1
6
7
7
9
6
4
5
1
2
8
3
4
2
3
7
9
6
8
1
5
1
5
9
8
4
2
7
6
3
6
8
7
1
2
3
9
5
8
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.