EQ-Bench Creative Writing v3

#	Model	Effort	Score	Cost / task	Harness
1	GPT-6 Astra	max	2173	–	–
2	Fable 5.1	max	2162	–	–
3	Opus 5	max	2133	–	–
4	GPT-6 Sol	max	2125	–	–
5	Kimi K3	max	2082	–	–
6	GLM 5.3	max	2075	–	–
7	Opus 5.5	max	2050	–	–
8	Grok 4.7	max	2007	–	–
9	GPT-5.6 Sol	max	1972	–	–
10	Fable 5	max	1943	–	–
11	Muse Spark 1.1	max	1927	–	–
12	Opus 4.7	max	1914	–	–
13	Muse Spark 1.3	max	1906	–	–
14	GPT-5.6 Terra	max	1855	–	–
15	GPT-5.5	xhigh	1844	–	–
16	Qwen 3.8	xhigh	1843	–	–
17	Muse Spark 1.2	max	1840	–	–
18	Opus 4.8	max	1840	–	–
19	GPT-5.4	xhigh	1840	–	–
20	GPT-5.6 Luna	max	1829	–	–
21	Sonnet 4.6	max	1810	–	–
22	Opus 4.6	max	1809	–	–
23	Sonnet 5	max	1794	–	–
24	GLM 5.2	max	1757	–	–
25	Gemini 3.8 Flash	max	1748	–	–
26	Kimi K2.6	max	1725	–	–
27	Gemini 3.7 Flash	max	1723	–	–
28	GPT-5.2	xhigh	1703	–	–
29	Nemotron 3 Ultra	max	1692	–	–
30	GPT-5.3	xhigh	1690	–	–
31	Opus 4.5	high	1687	–	–
32	Sonnet 4.5	high	1678	–	–
33	o3	high	1676	–	–
34	Qwen 3.8 27B	max	1671	–	–
35	Kimi K2	medium	1666	–	–
36	GPT-5.4 Mini	xhigh	1665	–	–
37	GPT-5.2025	high	1627	–	–
38	Inkling	max	1608	–	–
39	Gemini 3.6 Flash	xhigh	1604	–	–
40	GLM 5	high	1601	–	–
41	GLM 5.1	max	1592	–	–
42	Opus 4	medium	1580	–	–
43	Grok 4.5	max	1579	–	–
44	Kimi K2.5	max	1579	–	–
45	Grok 4.20	xhigh	1574	–	–
46	DeepSeek V4 Flash	max	1559	–	–
47	Gemini 3.5 Flash-Lite	xhigh	1559	–	–
48	DeepSeek V4 Pro	max	1553	–	–
49	DeepSeek V4.1 Flash	max	1540	–	–
50	Gemini 3 Pro	high	1525	–	–
51	DeepSeek V3.2	high	1515	–	–
52	DeepSeek R1	xhigh	1500	–	–
53	MiMo V2.5 Pro	xhigh	1493	–	–
54	Inkling Small	xhigh	1491	–	–
55	Gemini 3.1 Pro	high	1491	–	–
56	Sonnet 4	high	1483	–	–
57	Qwen 3.5 397B-A17B	max	1478	–	–
58	Mistral Medium 3.1	none	1476	–	–
59	DeepSeek V3.0324	–	1472	–	–
60	Sonnet 3.5	medium	1451	–	–
61	DeepSeek V4 Flash 0731	max	1438	–	–
62	DeepSeek V3.1	none	1436	–	–
63	Gemini 2.5 Pro	high	1421	–	–
64	DeepSeek R1.0528	–	1421	–	–
65	GPT-4.1	high	1420	–	–
66	GLM 4.7	none	1413	–	–
67	Mistral Large 3	none	1412	–	–
68	Sonnet 3.7	medium	1412	–	–
69	GLM 4.6	none	1411	–	–
70	Mistral Small	none	1369	–	–
71	Gemma 4 31B IT	high	1366	–	–
72	Qwen 3 235B-A22B	xhigh	1366	–	–
73	MiniMax M2.5	high	1361	–	–
74	GLM 4.5	xhigh	1341	–	–
75	Grok 4.1 Fast	high	1327	–	–
76	GPT-5 Mini	high	1313	–	–
77	Gemma 4 26B	xhigh	1305	–	–
78	Gemma 4 12B IT	–	1289	–	–
79	Nemotron 3.5 Lightning	–	1280	–	–
80	Gemma 3 27B IT	medium	1266	–	–
81	GPT-4.5	none	1258	–	–
82	Mistral Small 3.2	none	1255	–	–
83	Grok 3	none	1186	–	–
84	GPT-4.1 Mini	high	1147	–	–
85	Gemini 2.5 Flash	medium	1137	–	–
86	Gemini 2.0 Flash	medium	1128	–	–
87	Gemma 3 12B IT	–	1126	–	–
88	GLM 4.7 Flash	none	1125	–	–
89	GLM 4	–	1068	–	–
90	Gemma 3 4B IT	–	1068	–	–
91	Mistral Large 2411	–	985	–	–
92	gpt-oss-120b	high	961	–	–
93	GPT-4.1 Nano	high	946	–	–
94	Mistral	none	881	–	–
95	GPT-4o mini	high	873	–	–
96	Llama 3.1 405b	none	870	–	–
97	Llama 4 Maverick	medium	860	–	–
98	Gemma 2 9B IT	–	841	–	–
99	Llama 3.1 70b	none	784	–	–
100	Llama 4 Scout	medium	783	–	–
101	Mistral Small 3.1	none	761	–	–
102	GPT-4.0314	–	752	–	–
103	Haiku 3	medium	717	–	–
104	Llama 3.1 8b	medium	713	–	–
105	GPT-5 Nano	high	705	–	–
106	gpt-oss-20b	high	666	–	–
107	Llama 3.2 3b	none	595	–	–
108	GPT-3.5	high	451	–	–
109	Llama 3.2 1b	none	200	–	–