Benchmarks / Knowledge

UGI NatInt

A private-question evaluation of general knowledge and reasoning across textbook, popular-culture and world-model topics. Maintained by DontPlanToEnd. Score: NatInt 💡.

166models

–Open Index weight

–last updated

benchmarkheavenpublisher

#	Model	Effort	Score	Cost / task	Harness
1	GPT-6 Astra	max	79	–	–
2	Gemini 3.1 Pro	high	76	–	–
3	Gemini 3.7 Flash	max	76	–	–
4	GPT-6 Sol	max	74	–	–
5	Gemini 3.5 Flash	xhigh	72	–	–
6	GPT-6.1 Sol	max	71	–	–
7	Gemini 3 Flash	high	70	–	–
8	Opus 5.5	max	70	–	–
9	GPT-5.6 Sol	max	70	–	–
10	Gemini 3.8 Flash	max	70	–	–
11	Fable 5	max	68	–	–
12	Gemini 3 Pro	high	67	–	–
13	GPT-5.2025	high	67	–	–
14	Gemini 3.6 Flash	xhigh	66	–	–
15	o3	high	66	–	–
16	Grok 4.0709	–	65	–	–
17	Fable 5.1	max	65	–	–
18	o1	high	64	–	–
19	Gemini 2.5 Pro	high	64	–	–
20	Opus 4.5	high	64	–	–
21	GPT-5.3	xhigh	64	–	–
22	Opus 4.6	max	63	–	–
23	GPT-5.5	xhigh	63	–	–
24	Grok 4.6	max	61	–	–
25	Opus 4.8	max	61	–	–
26	Kimi K2.5	max	60	–	–
27	Grok 4.5	max	60	–	–
28	GPT-4.1	high	60	–	–
29	Kimi K2.6	max	59	–	–
30	DeepSeek V4 Pro	max	59	–	–
31	GPT-4o	high	59	–	–
32	Grok 4.7	max	59	–	–
33	GPT-5.1	high	58	–	–
34	Opus 4.1	high	58	–	–
35	Opus 4	medium	58	–	–
36	GPT-5.4	xhigh	57	–	–
37	GPT-5	high	57	–	–
38	Grok 3	none	57	–	–
39	Grok 4.20	xhigh	56	–	–
40	GPT-5.2	xhigh	56	–	–
41	GLM 5.2	max	55	–	–
42	Opus 4.7	max	55	–	–
43	Grok 4.3	xhigh	54	–	–
44	Sonnet 4.5	high	54	–	–
45	Sonnet 4.6	max	53	–	–
46	GPT-5.6 Terra	max	53	–	–
47	DeepSeek R1	xhigh	52	–	–
48	Sonnet 4	high	51	–	–
49	Opus 3	none	50	–	–
50	GLM 5.1	max	50	–	–
51	DeepSeek V3.1	none	49	–	–
52	GLM 5	high	49	–	–
53	Kimi K2	medium	49	–	–
54	DeepSeek V4 Flash	max	48	–	–
55	DeepSeek R1.0528	–	48	–	–
56	Qwen 3.5 397B-A17B	max	48	–	–
57	DeepSeek V3.0324	–	47	–	–
58	GLM 4.5	xhigh	47	–	–
59	DeepSeek V3.2	high	46	–	–
60	GPT-5.6 Luna	max	46	–	–
61	DeepSeek V3	none	46	–	–
62	o4 mini	high	46	–	–
63	GLM 4.6	none	45	–	–
64	Qwen 3 Max	max	44	–	–
65	Gemini 2.5 Flash	medium	44	–	–
66	Llama 3.1 405b	none	43	–	–
67	GLM 4.7	none	43	–	–
68	Qwen 3.6 Plus	max	42	–	–
69	Mistral Medium 3.5	high	41	–	–
70	MiMo V2.5 Pro	xhigh	41	–	–
71	Sonnet 5	max	40	–	–
72	Mistral Large 3	none	40	–	–
73	Mistral Medium 3.1	none	40	–	–
74	Gemma 4 31B IT	high	39	–	–
75	MiniMax M2.5	high	39	–	–
76	Mistral Large 2407	–	38	–	–
77	Hy3	xhigh	38	–	–
78	Qwen 3 235B-A22B	xhigh	36	–	–
79	Mistral Large 2411	–	36	–	–
80	Qwen 3.5 27B	high	36	–	–
81	Haiku 4.5	xhigh	35	–	–
82	Mistral Medium 3	none	35	–	–
83	Gemma 3 27B IT	medium	34	–	–
84	Gemma 4 26B	xhigh	34	–	–
85	MiniMax M2.7	high	34	–	–
86	MiniMax M2.1	–	34	–	–
87	Gemma 3 27B	none	34	–	–
88	gpt-oss-120b	high	34	–	–
89	Qwen 3.6 27B	max	33	–	–
90	Grok 4 Fast	none	33	–	–
91	Llama 3	–	33	–	–
92	GLM 4.6 V	low	32	–	–
93	MiMo V2.5	xhigh	32	–	–
94	Llama 33	–	32	–	–
95	MiniMax M2	–	32	–	–
96	Llama 33 70b	–	31	–	–
97	Grok 4.1 Fast	high	31	–	–
98	Llama 3.3 70b	none	31	–	–
99	Qwen 3.5 122B-A10B	none	31	–	–
100	Llama 4 Maverick	medium	31	–	–
101	Llama 3.1 70b	none	31	–	–
102	MiMo V2 Flash	high	31	–	–
103	Llama 3.3	none	30	–	–
104	Llama 7 0b	–	30	–	–
105	Haiku 3	medium	30	–	–
106	Gemma 2 7B	–	29	–	–
107	Llama 10 9b	–	28	–	–
108	Mistral Small 4	high	28	–	–
109	Mistral Small 3.1	none	27	–	–
110	Mistral Small	none	26	–	–
111	Gemma 4 12B IT	–	26	–	–
112	Qwen 3.6 35B-A3B	max	25	–	–
113	Mistral Small 3	none	25	–	–
114	Mistral Small 2409	–	25	–	–
115	Qwen 3.5 35B-A3B	low	25	–	–
116	Gemma 2 27B IT	–	25	–	–
117	Magistral 2509	–	24	–	–
118	Magistral Small 2509	–	24	–	–
119	Mistral Small 3.2	none	24	–	–
120	GLM 4	–	23	–	–
121	Qwen 2.5 32B	none	22	–	–
122	Devstral Small 2	none	21	–	–
123	Llama 4 Scout	medium	21	–	–
124	Mistral	none	21	–	–
125	Qwen 3 14B	none	21	–	–
126	Qwen 3 32B	none	20	–	–
127	GLM 4.7 Flash	none	20	–	–
128	Qwen 2.5 14B	–	19	–	–
129	Qwen 2.5 72B	none	19	–	–
130	Gemma 3 12B IT	–	19	–	–
131	GLM 4.5 Air	–	18	–	–
132	Llama 2 70b	none	18	–	–
133	Llama 3.1 8b	medium	18	–	–
134	Qwen 3 Coder	medium	18	–	–
135	Qwen 3.5 9B	low	18	–	–
136	gpt-oss-20b	high	17	–	–
137	Gemma 3 12B	none	17	–	–
138	Gemma 1 2B	–	17	–	–
139	Llama 3.3 8b	–	16	–	–
140	Gemma 2 9B	–	16	–	–
141	Qwen 3 30B-A3B	none	16	–	–
142	Qwen 2.5 7B	–	16	–	–
143	Qwen 3 8B	none	15	–	–
144	Qwen 3 15B-A2B	–	15	–	–
145	Qwen 2.5 Coder	none	14	–	–
146	Qwen 3.5 4B	low	13	–	–
147	Qwen 3	high	13	–	–
148	Gemma 2 9B IT	–	13	–	–
149	Qwen 3.1	none	12	–	–
150	Qwen 3 4B	none	12	–	–
151	Llama 3 8b	none	12	–	–
152	Gemma 3 4B IT	–	11	–	–
153	Llama 3 3b	–	11	–	–
154	Llama 3.2 3b	none	10	–	–
155	Gemma 3 4B	none	10	–	–
156	Gemma 2 2B IT	–	9	–	–
157	Qwen 3.5 2B	low	9	–	–
158	Qwen 2.5	–	9	–	–
159	Llama 3.1	max	9	–	–
160	Qwen 3 16B-A3B	–	8	–	–
161	Llama 3.2 1b	none	7	–	–
162	Llama 3.2	–	5	–	–
163	Qwen 3.0	–	5	–	–
164	Qwen 3.5	none	5	–	–
165	Gemma 3 1B IT	–	5	–	–
166	Gemma 3 1B	none	3	–	–