MineBench · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	Gemini 4	max	2253	–	–
2	GPT-6 Astra Pro	max	2250	–	–
3	GPT-6.1 Sol Pro	–	2188	–	–
4	Opus 5.5	max	2163	–	–
5	GPT-6 Sol Pro	max	2083	–	–
6	Opus 5	max	2027	–	–
7	Sonnet 5.5	max	2018	–	–
8	GPT-5.6 Sol Pro	max	2012	–	–
9	Fable 5.1	max	1946	–	–
10	GPT-5.5 Pro	xhigh	1938	–	–
11	Muse Wonder	max	1885	–	–
12	Fable 5	max	1885	–	–
13	GPT-5.5	xhigh	1862	–	–
14	DeepSeek V4.1 Flash	max	1839	–	–
15	Grok 4.6	max	1830	–	–
16	Gemini 3.8 Flash	max	1820	–	–
17	GPT-5.4 Pro	xhigh	1809	–	–
18	Grok 4.7	max	1806	–	–
19	Opus 4.8	max	1800	–	–
20	GLM 5.3 Flash	max	1798	–	–
21	Gemini 3.7 Flash	max	1780	–	–
22	GLM 5.3	max	1777	–	–
23	GPT-6 Luna Pro	max	1763	–	–
24	Muse Spark 1.3	max	1761	–	–
25	Kimi K3	max	1725	–	–
26	GPT-5.6 Luna Pro	–	1719	–	–
27	Gemini 3.1 Pro	high	1691	–	–
28	GPT-5.4	xhigh	1681	–	–
29	Gemini 3.5 Flash	xhigh	1667	–	–
30	Sonnet 5	max	1641	–	–
31	Gemini 3.6 Flash	xhigh	1635	–	–
32	Opus 4.7	max	1623	–	–
33	GPT-5.3 Codex	xhigh	1593	–	–
34	Muse Spark 1.2	max	1583	–	–
35	GLM 5.2	max	1577	–	–
36	GPT-5.2 Pro	xhigh	1569	–	–
37	Grok 4.5	max	1568	–	–
38	GLM 5.1	max	1530	–	–
39	DeepSeek V4 Flash 0731	max	1513	–	–
40	Opus 4.6	max	1502	–	–
41	GPT-5.2	xhigh	1486	–	–
42	DeepSeek V4 Pro	max	1469	–	–
43	Kimi K2.6	max	1455	–	–
44	GPT-5.4 Mini	xhigh	1416	–	–
45	Gemini 3 Pro	high	1404	–	–
46	Sonnet 4.6	max	1373	–	–
47	Gemini 3.5 Flash-Lite	xhigh	1366	–	–
48	Opus 4.5	high	1336	–	–
49	Grok 4.20	xhigh	1309	–	–
50	Qwen 3.5 397B-A17B	max	1304	–	–
51	GPT-5.4 Nano	xhigh	1276	–	–
52	Gemini 3 Flash	high	1272	–	–
53	Qwen 3.8 Max	max	1243	–	–
54	GLM 5	high	1243	–	–
55	Gemma 4 31B	max	1240	–	–
56	MiniMax M2.7	high	1222	–	–
57	Kimi K2.5	max	1215	–	–
58	Grok 4.3	xhigh	1176	–	–
59	Sonnet 4.5	high	1167	–	–
60	Gemini 2.5 Pro	high	1155	–	–
61	GLM 4.7	none	1128	–	–
62	GPT-5 Mini	high	1104	–	–
63	MiniMax M2.5	high	1096	–	–
64	GPT-5.2 Codex	xhigh	1081	–	–
65	DeepSeek V3.2	high	1065	–	–
66	GPT-4.1	high	1060	–	–
67	Qwen 3 Max	max	1018	–	–
68	Kimi K2	medium	990	–	–
69	Grok 4.1 Fast	high	981	–	–
70	Gemini 3.1 Flash-Lite	high	972	–	–
71	GPT-4.5	none	942	–	–
72	gpt-oss-120b	high	806	–	–
73	GPT-4o	high	773	–	–
74	Llama 4 Maverick	medium	656	–	–
75	GPT-5 Nano	high	589	–	–