Released: 2026-02-13
openai/gpt-4o
openai
7
1
4
8
9
5
3
6
2
5
6
2
7
3
4
1
9
8
3
8
9
2
6
1
4
7
5
8
2
6
9
4
3
5
1
7
9
5
4
2
7
6
8
3
1
1
7
3
5
8
8
6
2
4
6
9
8
1
2
7
5
3
4
4
3
1
7
5
9
2
8
6
2
4
7
6
1
3
9
5
9
mistralai/mistral-medium-3
mistral
7
5
6
8
9
4
3
2
1
1
9
2
7
3
5
6
8
4
3
8
9
2
6
1
4
7
5
8
7
6
9
4
3
5
1
2
5
2
4
2
5
7
8
9
1
9
4
3
1
2
8
7
6
5
2
1
8
4
2
7
9
3
6
6
3
1
5
8
9
2
4
7
4
6
5
6
1
2
7
5
9
anthropic/claude-3.7-sonnet
anthropic
7
1
5
4
9
6
3
2
8
3
6
2
7
8
5
1
9
4
9
8
7
3
5
1
4
6
2
1
5
6
9
4
3
2
8
7
8
9
4
2
7
5
6
1
3
2
7
3
1
6
8
5
4
9
4
5
8
9
2
7
1
3
6
6
3
1
5
4
9
8
7
2
7
2
4
6
1
3
8
5
9
deepseek/deepseek-chat-v3-0324
deepseek
7
1
5
4
9
2
3
6
8
4
6
2
7
3
5
1
9
8
8
9
3
6
5
1
4
2
7
3
2
6
9
4
1
5
8
7
5
7
4
2
6
3
9
8
1
1
9
7
5
7
8
2
4
6
9
5
8
1
2
7
6
3
4
6
3
1
8
5
4
7
2
9
2
4
7
6
1
3
8
5
9
x-ai/grok-3-mini-beta
x-ai
7
8
5
4
9
6
3
1
2
9
1
2
7
3
5
6
4
8
3
4
6
8
1
2
7
9
5
1
2
6
9
4
7
5
8
3
5
9
4
2
6
3
8
7
1
6
7
3
1
5
8
9
2
4
4
6
8
5
2
7
1
3
9
2
3
1
6
8
9
4
5
7
8
7
9
3
5
4
2
6
1
meta-llama/llama-3.3-70b-instruct
meta
7
5
1
8
9
6
3
2
4
2
9
6
7
5
3
1
8
5
8
1
3
4
7
1
4
6
2
5
4
6
9
4
2
8
7
1
1
8
4
2
6
5
7
9
1
9
6
5
1
8
8
2
4
3
3
2
8
5
2
7
6
3
9
6
3
1
5
3
4
9
1
8
4
7
9
6
1
3
5
5
9
google/gemini-2.0-flash-001
7
4
5
8
9
6
3
1
2
8
1
2
7
3
4
5
9
6
3
6
9
5
1
2
4
7
8
1
8
6
9
4
5
2
3
7
9
5
4
2
7
3
6
8
1
2
7
3
1
6
8
9
4
5
6
9
8
4
2
7
1
5
3
5
3
1
2
8
9
7
6
4
4
2
7
6
5
1
8
3
9
qwen/qwen3-235b-a22b
qwen
7
4
5
1
9
6
3
8
2
1
6
2
7
8
3
5
9
4
9
8
3
5
7
1
4
6
2
3
1
6
9
4
5
8
2
7
5
9
4
2
6
3
7
1
8
2
7
1
4
5
8
9
3
6
6
5
8
9
2
7
1
3
4
4
3
9
8
1
5
2
7
6
8
2
7
6
3
4
1
5
9
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.