Vals Index · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	Gemini 4	high	68.9%	$15.68	–
2	Sonnet 5.5	max	67.0%	$21.34	–
3	Opus 5.5	max	67.0%	$32.14	–
4	Fable 5.1	max	65.8%	$28.71	–
5	Wonder Full	max	64.6%	$8.56	–
6	Opus 5	max	63.7%	$19.31	–
7	GPT-6 Astra	max	63.1%	$18.46	–
8	Fable 5	max	61.4%	$29.59	–
9	GPT-6.1 Sol	max	61.2%	$3.24	–
10	Wonder	max	61.1%	$11.43	–
11	Muse Spark 1.3	max	58.2%	$3.79	–
12	GPT-5.6 Sol	max	58.0%	$14.24	–
13	GPT-6 Sol	max	57.5%	$7.58	–
14	MiMo V2.6 Pro	high	55.2%	$0.410	–
15	Opus 4.8	max	55.1%	$13.14	–
16	Grok 4.7	xhigh	54.9%	$12.11	–
17	Gemini 3.8 Flash	high	54.9%	$5.71	–
18	Haiku 5.5	max	54.4%	$3.00	–
19	GLM 5.3	max	53.5%	$7.25	–
20	MiMo V2.6 Flash	high	53.2%	$0.202	–
21	GPT-5.6 Terra	max	53.1%	$5.78	–
22	Grok 4.6	high	52.1%	$4.49	–
23	Sonnet 5	max	51.9%	$13.69	–
24	GPT-5.6 Luna	max	51.7%	$0.824	–
25	DeepSeek V4.1 Flash	high	51.3%	$0.332	–
26	Gemini 3.7 Flash	high	51.3%	$4.67	–
27	GPT-6 Luna	max	51.2%	$0.431	–
28	Ember-1	–	50.9%	$9.66	–
29	Kimi K3	max	50.3%	$6.44	–
30	Hy4	–	49.9%	$1.44	–
31	Step 5 Preview	–	49.4%	$2.63	–
32	Muse Spark 1.2	xhigh	49.3%	$1.94	–
33	Qwen 3.8 Max	max	48.3%	$4.96	–
34	Mistral Large 4	high	48.1%	$13.78	–
35	DeepSeek V4 Flash 0731	high	48.0%	$0.848	–
36	DeepSeek V4 Pro 0813	max	47.6%	$3.35	–
37	Gemini 3.5 Flash	high	44.8%	$3.30	–
38	Grok 4.5	high	44.7%	$2.42	–
39	GLM 5.3 Flash	max	44.6%	$1.03	–
40	DeepSeek V4	max	38.6%	$1.31	–
41	MiniMax M3	max	36.5%	$3.08	–
42	MiMo V2.5 Pro	xhigh	33.8%	$0.162	–
43	Gemini 3.1 Pro	high	33.4%	$2.18	–
44	GPT-5.4 Mini	xhigh	33.2%	$1.45	–
45	Inkling	max	28.7%	$1.46	–
46	Inkling Small	xhigh	25.5%	$0.282	–
47	Mercury 2.5	high	8.9%	$0.285	–