WeirdML v2 · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	GPT-6 Astra Pro	xhigh	93.6%	$29.31	–
2	GPT-6 Astra	xhigh	93.3%	$7.75	–
3	Fable 5.1	xhigh	92.9%	$11.69	–
4	Fable 5	xhigh	91.9%	$8.42	–
5	Opus 5	xhigh	91.8%	$4.74	–
6	GPT-5.6 Sol Pro	xhigh	89.4%	$3.87	–
7	GPT-5.6 Sol	max	88.8%	$1.25	–
8	GPT-5.5	xhigh	84.9%	$2.57	–
9	Gemini 3.8 Flash	max	84.8%	$0.507	–
10	Opus 4.8	xhigh	82.9%	$2.35	–
11	Kimi K3	max	82.6%	$2.05	–
12	GPT-5.6 Terra	max	78.3%	$0.250	–
13	Opus 4.6	max	77.9%	$5.78	–
14	GPT-5.3 Codex	xhigh	77.9%	$2.45	–
15	GPT-5.4	xhigh	77.7%	$5.72	–
16	Opus 4.7	max	76.4%	$1.82	–
17	GLM 5.3	max	75.4%	$0.535	–
18	Qwen 3.8	xhigh	75.2%	$1.42	–
19	GPT-5.2	xhigh	72.2%	$2.05	–
20	Gemini 3.1 Pro	high	72.1%	$1.36	–
21	GLM 5.2	xhigh	70.1%	$0.507	–
22	Gemini 3 Pro	high	69.9%	$0.526	–
23	Sonnet 5	high	68.8%	$0.895	–
24	Grok 4.6	max	67.3%	$0.335	–
25	DeepSeek V4 Pro 0813	max	66.2%	$0.191	–
26	Sonnet 4.6	max	66.1%	$2.86	–
27	Opus 4.5	high	63.7%	$0.738	–
28	DeepSeek V4 Flash 0731	max	63.0%	$0.029	–
29	Gemini 3.5 Flash	xhigh	62.6%	$0.756	–
30	Gemini 3 Flash	high	61.6%	$0.222	–
31	GPT-5.6 Luna	max	60.9%	$0.040	–
32	GPT-5.1	high	60.8%	$0.691	–
33	GPT-5.2025	high	60.7%	$0.593	–
34	GPT-5 Pro	high	60.4%	$7.65	–
35	GPT-5.4 Mini	xhigh	60.3%	$1.29	–
36	Muse Spark 1.2	xhigh	60.3%	$0.380	–
37	o3 pro	high	58.2%	$2.54	–
38	GLM 5.1	max	57.1%	$0.318	–
39	Gemini 3.6 Flash	xhigh	56.1%	$0.623	–
40	Kimi K2.6	max	55.9%	$0.350	–
41	GPT-5 Codex	high	54.5%	$1.10	–
42	Kimi K2.7 Code	max	54.1%	$0.232	–
43	Gemini 2.5 Pro	high	54.0%	$0.410	–
44	GPT-5 Mini	high	52.7%	$0.110	–
45	o4 mini	high	52.6%	$0.196	–
46	o3	high	52.4%	$0.226	–
47	Grok 4.20	xhigh	52.3%	$0.126	–
48	Gemma 4 31B IT	high	52.3%	$0.013	–
49	Gemini 3.1 Flash-Lite	high	52.2%	$0.066	–
50	Grok 4.3	xhigh	49.9%	$0.090	–
51	GPT-5.4 Nano	xhigh	49.2%	$0.048	–
52	DeepSeek V4 Pro	max	48.9%	$0.085	–
53	GLM 5	high	48.2%	$0.079	–
54	gpt-oss-120b	high	48.2%	$0.033	–
55	o1	high	47.6%	$1.85	–
56	DeepSeek V3.2	high	46.7%	$0.035	–
57	Grok 4.5	max	46.4%	$0.080	–
58	Opus 4.1	high	45.9%	$1.76	–
59	Grok 4.07	–	45.7%	$0.452	–
60	DeepSeek V4 Flash	max	45.6%	$0.015	–
61	Kimi K2.5	max	45.6%	$0.053	–
62	Mistral Medium 3.5	high	43.7%	$0.278	–
63	o3 mini	high	43.7%	$0.174	–
64	Nemotron 3 Ultra	max	43.5%	$0.077	–
65	Mercury 2	high	43.2%	$0.030	–
66	Grok 4 Fast	none	42.9%	$0.013	–
67	Kimi K2	medium	42.8%	$0.081	–
68	Grok 3 Mini	high	42.6%	$0.017	–
69	Gemini 2.5 Flash	medium	41.9%	$0.073	–
70	DeepSeek R1.0528	–	41.6%	$0.071	–
71	Qwen 3 Coder	medium	41.2%	$0.073	–
72	Qwen 3 235B-A22B	xhigh	41.0%	$0.020	–
73	gpt-oss-20b	high	40.9%	$0.011	–
74	GLM 4.5	xhigh	40.6%	$0.073	–
75	Sonnet 3.5	medium	40.0%	$0.248	–
76	GPT-5	high	39.8%	$0.079	–
77	Qwen 3.5 27B	high	39.5%	$0.040	–
78	GPT-4.5	none	39.4%	$1.58	–
79	GPT-4.1	high	39.0%	$0.097	–
80	Gemini 3.5 Flash-Lite	xhigh	39.0%	$0.047	–
81	DeepSeek V3.1	none	38.4%	$0.031	–
82	GPT-5 Nano	high	38.1%	$0.026	–
83	GPT-4.1 Mini	high	37.6%	$0.017	–
84	Grok 3	none	37.2%	$0.248	–
85	MiniMax M2.7	high	36.9%	$0.035	–
86	Kimi K2.0905	–	36.7%	$0.042	–
87	DeepSeek R1	xhigh	36.5%	$0.079	–
88	o1 mini	high	36.3%	$0.111	–
89	DeepSeek V3.0324	–	36.1%	$0.014	–
90	Gemini 2.5 Flash Lite	none	35.2%	$0.030	–
91	Gemma 4 26B	xhigh	35.2%	$0.021	–
92	Qwen 3.6 35B-A3B	max	34.5%	$0.060	–
93	Mistral Medium 3.1	none	33.1%	$0.040	–
94	Inkling	max	32.3%	$0.126	–
95	Haiku 3.5	medium	30.7%	$0.057	–
96	Qwen 3 30B-A3B	none	29.8%	$0.0081	–
97	Gemini 2.0 Flash	medium	25.8%	$0.0067	–
98	GPT-4o	high	25.1%	$0.084	–
99	Llama 4 Maverick	medium	24.5%	$0.0061	–
100	Grok 2.1212	–	22.2%	$0.114	–
101	Llama 3.1 405b	none	21.4%	$0.020	–
102	GPT-4.1 Nano	high	19.0%	$0.0026	–
103	GPT-4	high	18.0%	$0.321	–
104	Qwen 2.5 72B	none	16.0%	$0.011	–
105	Llama 3.3 70b	none	14.4%	$0.0015	–
106	GPT-4o mini	high	11.8%	$0.0053	–
107	Qwen 2 72B	none	11.3%	$0.010	–
108	Haiku 3	medium	9.8%	$0.014	–
109	Llama 3.1 70b	none	9.0%	$0.0032	–
110	GPT-3.5	high	3.5%	$0.042	–
111	Llama 3.1 8b	medium	1.7%	$0.0004	–