feat: add results for gemini-3.1-pro, gpt-5.5, opus-4.6

[Abhinavexist]

No description provided.

[github-actions]

🏆 Leaderboard preview

Built 24 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	GLM-4.7	0.861	0.804	0.965	0.508
4	Qwen3.5-35B	0.861	0.801	0.969	0.500
5	gpt-5.5	0.860	0.795	0.978	0.464
6	Gemini-2.5-Flash	0.860	0.796	0.972	0.498
7	Qwen3-235B	0.857	0.786	0.978	0.463
8	Interfaze-Beta	0.855	0.795	0.967	0.480
9	Claude-Sonnet-4.6	0.854	0.779	0.979	0.442
10	claude-opus-4-6	0.853	0.779	0.977	0.437

Generated at 2026-04-29T23:50:28+00:00 • full JSON in workflow artifacts