feat(eval): add grok-4.3, gemini-2.5-pro, gpt-5.4-mini scores

[Abhinavexist]

No description provided.

[github-actions]

🏆 Leaderboard preview

Built 31 models, top 10 by Overall:

Rank	Model	Overall	Val. Acc.	JSON Pass	Perfect
1	GPT-5.4	0.870	0.798	0.993	0.469
2	gemini-3.1-pro-preview	0.869	0.820	0.966	0.542
3	z-ai_glm-5.1	0.866	0.806	0.975	0.498
4	claude-opus-4-7	0.864	0.787	0.993	0.424
5	GLM-4.7	0.861	0.804	0.965	0.508
6	Qwen3.5-35B	0.861	0.801	0.969	0.500
7	gpt-5.5	0.860	0.795	0.978	0.464
8	Gemini-2.5-Flash	0.860	0.796	0.972	0.498
9	Qwen3-235B	0.857	0.786	0.978	0.463
10	gemini-2.5-pro	0.856	0.805	0.958	0.504

Generated at 2026-05-06T18:11:16+00:00 • full JSON in workflow artifacts