Released: 2026-03-06
openai/gpt-4o
openai
2
9
8
1
5
7
6
4
3
7
6
5
4
3
9
2
8
1
1
3
4
2
6
8
9
5
7
8
1
6
9
2
3
7
4
5
4
7
5
8
9
6
1
3
2
3
2
9
7
4
5
8
6
1
6
4
7
3
1
2
5
9
8
5
8
3
6
7
5
1
9
4
9
5
2
6
8
1
3
7
6
deepseek/deepseek-chat-v3-0324
deepseek
2
1
8
5
6
7
9
4
3
6
5
4
9
3
2
2
8
1
7
3
9
8
2
4
6
5
1
3
2
6
9
8
1
7
4
5
4
7
5
2
9
6
8
1
2
8
9
1
7
4
3
3
6
9
9
4
7
3
1
8
5
2
6
5
8
3
4
7
5
1
9
2
1
6
2
6
5
9
4
3
8
google/gemini-3-flash-preview
2
5
8
1
9
7
6
4
3
6
7
4
5
3
9
2
8
1
9
3
1
4
2
6
7
5
8
3
1
6
9
8
2
7
4
5
4
9
5
7
6
3
8
1
2
8
7
2
5
4
1
3
6
9
8
6
7
3
1
4
5
2
9
2
8
3
4
7
5
1
9
6
1
4
9
6
5
2
8
3
7
mistralai/mistral-medium-3
mistral
2
4
9
8
5
7
6
3
1
7
5
6
4
3
9
2
8
1
8
3
1
2
6
7
9
5
4
3
7
6
9
8
1
7
4
5
4
9
5
3
7
2
8
1
6
1
2
8
5
4
6
3
9
7
6
2
7
8
1
3
5
4
9
5
8
3
7
9
5
1
6
2
9
1
4
6
2
8
7
3
5
google/gemini-2.0-flash-001
2
5
8
1
6
7
9
3
4
7
6
4
5
3
9
2
8
1
9
3
1
8
2
4
6
5
7
3
1
6
9
8
2
7
4
5
4
7
5
3
2
6
8
1
9
8
9
2
5
4
1
3
6
7
6
2
7
4
1
8
5
9
3
5
8
3
7
9
5
1
9
6
1
4
9
6
5
3
4
7
8
anthropic/claude-3.7-sonnet
anthropic
2
1
9
4
6
7
8
5
3
7
5
4
9
3
6
2
8
1
6
3
8
1
5
2
9
4
7
1
2
6
9
8
3
7
4
5
4
7
5
2
9
1
6
3
8
8
9
1
5
4
7
3
6
2
9
4
7
3
1
8
5
2
6
3
8
2
7
6
5
1
9
4
5
6
3
6
2
9
4
7
9
meta-llama/llama-3.3-70b-instruct
meta
2
1
4
8
5
7
3
6
9
5
9
3
1
3
2
2
8
1
8
3
2
5
9
6
1
4
7
1
5
6
9
7
3
7
2
5
4
2
5
3
1
8
9
7
6
3
7
1
4
4
5
3
6
8
9
6
7
2
1
4
5
3
2
6
8
9
7
2
5
4
9
1
7
4
3
6
8
9
1
5
2
x-ai/grok-3-mini-beta
x-ai
2
5
4
1
6
7
8
3
9
6
7
9
5
3
8
2
4
1
1
3
2
4
8
9
6
7
5
8
1
6
9
2
4
7
5
3
4
9
5
7
1
3
8
6
2
7
2
8
3
4
5
9
1
6
3
6
7
8
9
1
5
2
4
5
4
3
2
7
6
1
9
8
9
8
1
6
5
2
4
3
7
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.