SWE-bench Verified (Vals) · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	Opus 5	max	97.0%	$1.29	–
2	DeepSeek V4 Pro 0813	max	96.4%	$0.103	–
3	GPT-5.6 Sol	max	96.2%	$1.15	–
4	Grok 4.6	high	95.6%	$0.785	–
5	GPT-5.6 Terra	max	95.4%	$0.401	–
6	GLM 5.3	max	95.4%	$0.338	–
7	Fable 5	max	95.0%	$2.05	–
8	Kimi K3	max	93.4%	$0.760	–
9	GPT-5.6 Luna	max	93.0%	$0.043	–
10	GLM 5.3 Flash	max	92.0%	$0.018	–
11	DeepSeek V4 Flash 0731	high	88.8%	$0.0099	–
12	Opus 4.8	max	88.6%	$1.92	–
13	Grok 4.5	high	86.6%	$0.540	–
14	Muse Spark 1.2	xhigh	86.6%	$0.554	–
15	Qwen 3.8 27B	xhigh	86.0%	$1.08	–
16	Qwen 3.8 Max	max	85.6%	$1.13	–
17	GLM 5.2	max	82.8%	$0.710	–
18	GPT-5.5	xhigh	82.6%	$1.36	–
19	Inkling Small	xhigh	82.2%	$0.134	–
20	Muse Spark 1.1	xhigh	82.0%	$0.352	–
21	Opus 4.7	max	82.0%	$2.42	–
22	Gemini 3.7 Flash	high	80.8%	$1.44	–
23	Gemini 3.8 Flash	high	80.0%	$2.19	–
24	Composer 2.5	none	79.6%	–	Cursor CLI
25	Gemini 3.6 Flash	high	79.6%	$1.19	–
26	Sonnet 5	max	79.6%	$1.49	–
27	Gemini 3.5 Flash	high	78.8%	$0.954	–
28	Gemini 3.1 Pro	high	78.8%	$0.783	–
29	GPT-5.4	xhigh	78.2%	$0.801	–
30	Opus 4.6	max	78.2%	$1.22	–
31	Kimi K2.7 Code	max	78.2%	$0.265	–
32	GPT-5.3 Codex	xhigh	78.0%	$0.455	–
33	Inkling	max	77.6%	$0.586	–
34	Sonnet 4.6	max	77.4%	$1.30	–
35	DeepSeek V4	max	77.4%	$0.437	–
36	Opus 4.5	high	76.4%	$1.05	–
37	Gemini 3 Pro	high	76.4%	$0.772	–
38	GLM 5.1	max	76.4%	$0.457	–
39	Kimi K2.6	max	76.2%	$0.275	–
40	GPT-5.2	xhigh	75.8%	$1.45	–
41	Gemini 3 Flash	high	75.0%	$0.280	–
42	Gemini 3.5 Flash-Lite	high	75.0%	$0.309	–
43	MiniMax M3	max	75.0%	$0.416	–
44	MiniMax M2.1	–	74.8%	$0.439	–
45	Muse Spark	–	74.4%	$0.444	–
46	MiniMax M2.5	high	74.2%	$0.457	–
47	MiMo V2.5 Pro	xhigh	74.0%	$0.023	–
48	MiniMax M2.7	high	73.8%	$0.466	–
49	Qwen 3.6 Plus	max	73.4%	$0.869	–
50	GPT-5.4 Mini	xhigh	73.0%	$0.511	–
51	Qwen 3.6 Max	max	72.8%	$1.57	–
52	GPT-5.2 Codex	high	72.4%	$0.312	–
53	Grok 4.20	xhigh	72.2%	$0.516	–
54	Grok 4.3	high	71.4%	$0.223	–
55	GLM 5	high	71.4%	$0.449	–
56	Qwen 3.5 Plus	none	71.2%	$0.964	–
57	MiMo V2.5	xhigh	71.0%	$0.010	–
58	Kimi K2.5	max	70.0%	$0.168	–
59	Sonnet 4.5	high	70.0%	$0.807	–
60	Qwen 3.6 27B	max	70.0%	$0.914	–
61	GPT-5.4 Nano	high	69.8%	$0.098	–
62	GPT-5.1	high	69.8%	$0.713	–
63	GLM 4.7	none	69.4%	$0.204	–
64	GPT-5	high	69.0%	$0.270	–
65	Nemotron 3 Ultra 550B A55B	max	69.0%	–	–
66	Qwen 3.7 Max	max	68.8%	$3.23	–
67	DeepSeek V3.2	high	67.6%	$0.673	–
68	Haiku 4.5	xhigh	66.6%	$0.366	–
69	Mistral Medium 3.5	high	66.4%	$4.00	–
70	Ling 3.0 Flash	–	65.2%	$0.032	–
71	Qwen 3.5 Flash	max	64.4%	$0.263	–
72	Gemini 3.1 Flash-Lite	high	62.8%	$0.105	–
73	Devstral 2	none	62.8%	$0.747	–
74	GPT-5 Mini	high	60.8%	$0.050	–
75	Kimi K2	medium	60.2%	$0.366	–
76	Grok 4	high	57.8%	$1.42	–
77	Laguna M.1	–	57.6%	$0.520	–
78	Laguna XS.2	–	55.2%	$0.170	–
79	Gemini 2.5 Pro	high	54.4%	$0.352	–
80	Nemotron 3.5 Lightning	–	52.8%	$0.032	–
81	Grok 4 Fast	none	45.4%	$0.051	–
82	Grok 4.1 Fast	high	41.4%	$0.028	–
83	Mistral Large 3	none	41.4%	$0.322	–
84	gpt-oss-120b	high	33.6%	$0.172	–
85	Command A	none	7.8%	$0.210	–