Optimization Problem Solving on OptiVerse benchmark leaderboard · SOTA2 Research

Optimization Problem Solving on OptiVerse

58.04Mean Performance (MP)Gemini-3-Pro

Evaluation Results

Method	Links
Gemini-3-Pro2026.04		58.04	57.14	56.67	56.85	42.31	50.98	89	52.75	27	55.9
GPT-5.22026.04		55.86	57.14	50	57.53	50	54.9	91	50.75	25.33	55.2
Gemini-3-Flash2026.04		54.77	59.24	54.17	55.48	48.72	56.86	88.67	53.25	25.33	55.5
DeepSeek-V3.2-Reasoner2026.04		53.68	52.52	45	49.32	28.21	49.02	84.33	44.5	21.33	49.5
Qwen3-235B-Thinking2026.04		53.68	51.26	49.17	52.74	43.59	49.02	87.33	47	21.33	51.4
Gemini-2.5-Pro2026.04		53.68	49.16	50	47.95	38.46	43.14	87	43.75	20	49.6
Claude-4.5-Sonnet2026.04		53.41	52.52	46.67	47.95	34.62	45.1	83.67	45.25	21.67	49.7
o32026.04		52.04	53.78	45.83	56.85	39.74	45.1	86.67	47.25	20.67	51.1
DeepSeek-V3.2-Chat2026.04		51.23	47.48	45	43.15	21.79	35.29	79.67	39.25	19	45.3
o4-mini2026.04		50.95	54.2	48.33	55.48	42.31	47.06	87.67	46.75	20.67	51.2
Qwen3-30B-Thinking2026.04		49.59	46.64	46.67	46.58	30.77	43.14	80.67	40	20.33	46.3
GPT-OSS-120B2026.04		49.54	55.33	34.19	38.93	32.05	35.29	78.67	39.25	18.67	44.9
Gemini-2.5-Flash2026.04		46.05	46.22	48.33	48.63	48.72	54.9	82.33	42.75	18.67	47.4
Qwen3-235B-Instruct2026.04		44.41	47.06	45.83	41.1	42.31	41.18	78.33	39.75	16.67	44.4
Qwen3-8B-Thinking2026.04		40.33	43.7	41.67	39.73	25.64	37.25	73	33	16	39.9
Kimi-K22026.04		40.05	44.54	38.33	39.73	24.36	25.49	71.33	31.5	16.33	38.9
Qwen3-30B-Instruct2026.04		38.96	36.13	31.67	34.25	19.23	15.69	70.33	21.75	14	34
Qwen2.5-72B2026.04		31.61	27.31	20	15.75	7.69	13.73	48	16.5	10.33	24.1
Qwen3-Coder-30B2026.04		30.79	31.09	23.33	23.29	8.97	9.8	49.67	19.25	11.67	26.1
Ministral3-8B-Instruct2026.04		26.01	18.44	19.35	12.08	2.56	5.88	39.33	14.75	5.67	18.2
Qwen3-8B-Instruct2026.04		23.98	22.69	20.83	14.38	6.41	13.73	42.67	12.25	7.67	20
Internlm3-8B-instruct2026.04		11.44	7.56	5	8.22	0	3.92	18.67	3.75	3	8