Coding

SWE-Rebench leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

As of 2026-10-07, the highest measured score on SWE-Rebench is 65.3% by Claude Opus 4.6. 158 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Opus 4.6	65.3%	measured
2	GLM-5	62.8%	measured
3	GLM-5.1	62.7%	measured
4	Gemini 4 Argon	62.3%	estimated ± 6.5 pp, low confidence
5	Mistral Large 4	62.2%	estimated ± 6.5 pp, low confidence
6	Claude Opus 4.7	62.2%	estimated ± 6.5 pp, low confidence
7	GPT-5.5	62.2%	estimated ± 6.5 pp, low confidence
8	GPT-5.4	62.2%	estimated ± 6.5 pp, low confidence
9	Claude Opus 4.6 (Adaptive)	62.1%	estimated ± 6.5 pp, low confidence
10	Gemini 3.5 Flash	62.0%	estimated ± 6.5 pp, low confidence
11	GPT-5.4 mini	62.0%	estimated ± 6.5 pp, low confidence
12	Claude Fable 5	61.8%	estimated ± 3.6 pp, low confidence
13	Claude Mythos 5	61.8%	estimated ± 3.6 pp, low confidence
14	Claude Opus 4.7 (Adaptive)	61.8%	estimated ± 3.6 pp, low confidence
15	Claude Opus 4.8	61.8%	estimated ± 3.6 pp, low confidence
16	Claude Opus 5	61.8%	estimated ± 3.6 pp, low confidence
17	Claude Sonnet 5	61.8%	estimated ± 3.6 pp, low confidence
18	Ember-1	61.8%	estimated ± 3.6 pp, low confidence
19	Ornith-1.5-397B	61.8%	estimated ± 3.6 pp, low confidence
20	Ornith-1.0-397B	61.8%	estimated ± 3.6 pp, medium confidence
21	Claude Opus 4.5	61.8%	estimated ± 3.6 pp, medium confidence
22	Beam	61.8%	estimated ± 3.6 pp, medium confidence
23	DeepSeek V4 Pro 0813	61.8%	estimated ± 3.6 pp, medium confidence
24	MiniMax M3	61.8%	estimated ± 3.6 pp, medium confidence
25	Qwen3.7 Max	61.8%	estimated ± 3.6 pp, medium confidence
26	Inkling-Small	61.8%	estimated ± 3.6 pp, medium confidence
27	Kimi K2.6	61.8%	estimated ± 3.6 pp, medium confidence
28	GPT-5.2	61.8%	estimated ± 3.6 pp, medium confidence
29	DeepSeek V4 Flash 0731	61.7%	estimated ± 3.6 pp, medium confidence
30	Ornith-1.5-35B-A3B	61.7%	estimated ± 3.6 pp, medium confidence
31	Qwen3.6 Plus	61.7%	estimated ± 3.6 pp, medium confidence
32	BTL-4	61.6%	estimated ± 3.6 pp, medium confidence
33	dots3-note Preview	61.6%	estimated ± 3.6 pp, medium confidence
34	MiMo-V2-Pro	61.3%	estimated ± 3.6 pp, medium confidence
35	Apodex 1.1	61.0%	estimated ± 3.6 pp, medium confidence
36	Qwen3.7 Plus	61.0%	estimated ± 3.6 pp, medium confidence
37	GPT-5.2-Codex	61.0%	estimated ± 6.5 pp, low confidence
38	DeepSeek V3.2	60.9%	measured
39	Inkling	60.9%	estimated ± 3.6 pp, medium confidence
40	Mistral Medium 3.5 128B	60.9%	estimated ± 3.6 pp, medium confidence
41	Claude Sonnet 4.6	60.7%	measured
42	Muse Spark	60.6%	estimated ± 3.6 pp, medium confidence
43	Claude Sonnet 4.5	60.2%	estimated ± 3.6 pp, medium confidence
44	Qwen3.6-27B	60.2%	estimated ± 3.6 pp, medium confidence
45	Kimi K2.5 (Reasoning)	59.3%	estimated ± 3.6 pp, medium confidence
46	Grok 4.20	59.0%	estimated ± 3.6 pp, medium confidence
47	Gemini 3.1 Pro	58.9%	estimated ± 6.5 pp, low confidence
48	Qwen3.5-27B	58.9%	measured
49	GLM-4.7	58.7%	measured
50	Kimi K2.5	58.5%	measured
51	GPT-5.3 Codex	58.2%	measured
52	Apodex 1.1 Mini	58.1%	estimated ± 8.9 pp, low confidence
53	Claude Fable 5.1	58.1%	estimated ± 8.9 pp, low confidence
54	Gemini 3.6 Flash	58.1%	estimated ± 8.9 pp, low confidence
55	Gemini 3.7 Flash	58.1%	estimated ± 8.9 pp, low confidence
56	Gemini 3.8 Flash	58.1%	estimated ± 8.9 pp, low confidence
57	GLM-5.2	58.1%	estimated ± 8.9 pp, low confidence
58	GLM-5.3	58.1%	estimated ± 8.9 pp, low confidence
59	GPT-5.6 Luna	58.1%	estimated ± 8.9 pp, low confidence
60	GPT-5.6 Sol	58.1%	estimated ± 8.9 pp, low confidence
61	GPT-5.6 Terra	58.1%	estimated ± 8.9 pp, low confidence
62	GPT-6 Astra	58.1%	estimated ± 8.9 pp, low confidence
63	Grok 4.5	58.1%	estimated ± 8.9 pp, low confidence
64	Grok 4.6	58.1%	estimated ± 8.9 pp, low confidence
65	Hy3	58.1%	estimated ± 8.9 pp, low confidence
66	Kimi K2.7 Code	58.1%	estimated ± 8.9 pp, low confidence
67	Kimi K3	58.1%	estimated ± 8.9 pp, low confidence
68	Ling 3.0 Flash	58.1%	estimated ± 8.9 pp, low confidence
69	Ling 3.0 Flash FP8	58.1%	estimated ± 8.9 pp, low confidence
70	MiMo-V2.5-Pro	58.1%	estimated ± 8.9 pp, low confidence
71	Muse Spark 1.1	58.1%	estimated ± 8.9 pp, low confidence
72	Muse Spark 1.2	58.1%	estimated ± 8.9 pp, low confidence
73	Muse Spark 1.3	58.1%	estimated ± 8.9 pp, low confidence
74	Quasar 438B	58.1%	estimated ± 8.9 pp, low confidence
75	Qwen3.8-27B	58.1%	estimated ± 8.9 pp, low confidence
76	Qwen3.8-Flash-Next	58.1%	estimated ± 8.9 pp, low confidence
77	Qwen3.8 Max Preview	58.1%	estimated ± 8.9 pp, low confidence
78	Gemini 3.5 Flash-Lite	58.1%	estimated ± 8.9 pp, low confidence
79	Qwen3.5 397B	58.0%	estimated ± 3.6 pp, medium confidence
80	Composer 2	58.0%	measured
81	Muse Glimmer 30B	57.8%	estimated ± 3.6 pp, medium confidence
82	Ornith-1.0-35B	57.4%	estimated ± 3.6 pp, medium confidence
83	MiMo-V2-Omni	57.1%	estimated ± 3.6 pp, medium confidence
84	Laguna M.1	57.1%	estimated ± 3.6 pp, medium confidence
85	Claude 4.1 Opus	57.1%	estimated ± 3.6 pp, medium confidence
86	Hy3 Preview	57.1%	estimated ± 3.6 pp, medium confidence
87	MAI-Thinking-1	57.0%	estimated ± 3.6 pp, medium confidence
88	MiMo-V2-Flash	57.0%	estimated ± 3.6 pp, medium confidence
89	Qwen3.6-35B-A3B	57.0%	estimated ± 3.6 pp, medium confidence
90	Claude Haiku 4.5	57.0%	estimated ± 3.6 pp, medium confidence
91	Claude 4 Sonnet	57.0%	estimated ± 3.6 pp, medium confidence
92	MAI-Code-1.1-Flash	57.0%	estimated ± 3.6 pp, medium confidence
93	Nemotron 3 Ultra	57.0%	estimated ± 3.6 pp, medium confidence
94	Qwen3.5-122B-A10B	57.0%	estimated ± 3.6 pp, medium confidence
95	Claude 3.5 Sonnet	57.0%	estimated ± 3.6 pp, low confidence
96	DeepSeek V3	57.0%	estimated ± 3.6 pp, low confidence
97	Gemini 2.5 Pro	57.0%	estimated ± 3.6 pp, low confidence
98	GPT-4.1	57.0%	estimated ± 3.6 pp, low confidence
99	GPT-4.1 mini	57.0%	estimated ± 3.6 pp, low confidence
100	Granite 4.2 30B	57.0%	estimated ± 3.6 pp, low confidence
101	Granite 4.2 8B	57.0%	estimated ± 3.6 pp, low confidence
102	Grok Code Fast 1	57.0%	estimated ± 3.6 pp, medium confidence
103	K-EXAONE 2.0	57.0%	estimated ± 3.6 pp, low confidence
104	Laguna XS.2	57.0%	estimated ± 3.6 pp, medium confidence
105	Laguna XS 2.1	57.0%	estimated ± 3.6 pp, medium confidence
106	LLaDA2.2-flash	57.0%	estimated ± 3.6 pp, low confidence
107	LongCat-Flash-Lite-Sparse	57.0%	estimated ± 3.6 pp, low confidence
108	MiniCPM5-2B	57.0%	estimated ± 3.6 pp, low confidence
109	Nemotron 3.5 Lightning 30B A3B NVFP4	57.0%	estimated ± 3.6 pp, low confidence
110	o3-mini	57.0%	estimated ± 3.6 pp, low confidence
111	Ornith-1.0-9B	57.0%	estimated ± 3.6 pp, medium confidence
112	Ornith-1.5-9B	57.0%	estimated ± 3.6 pp, medium confidence
113	Solar Open 2	57.0%	estimated ± 3.6 pp, medium confidence
114	Solar Pro 4	57.0%	estimated ± 3.6 pp, medium confidence
115	Ternary Bonsai 2 27B	57.0%	estimated ± 3.6 pp, low confidence
116	ZAYA1-74B-Preview	57.0%	estimated ± 3.6 pp, low confidence
117	Qwen3.5-35B-A3B	53.7%	measured
118	GPT-5.4 nano	52.2%	estimated ± 6.5 pp, low confidence
119	MiniMax M2.7	51.9%	measured
120	GPT-5.1	48.8%	estimated ± 6.5 pp, low confidence
121	GLM-5 (Reasoning)	45.2%	estimated ± 6.5 pp, low confidence
122	Claude Sonnet 4.5 Thinking	42.8%	estimated ± 6.5 pp, low confidence
123	Gemma 4 31B	41.6%	measured
124	GPT-5.1-Codex-Max	41.1%	estimated ± 6.5 pp, low confidence
125	Claude Opus 4.5 Thinking	34.7%	estimated ± 6.5 pp, low confidence
126	Gemini 3 Flash	32.8%	estimated ± 6.5 pp, low confidence
127	GPT-5 (high)	32.3%	estimated ± 6.5 pp, low confidence
128	Qwen3.5 Plus	12.4%	estimated ± 6.5 pp, low confidence
129	Grok 4.3	11.2%	estimated ± 8.9 pp, low confidence
130	MiniMax M2.5	9.3%	estimated ± 6.5 pp, low confidence
131	Gemini 3 Pro	7.6%	estimated ± 6.5 pp, low confidence
132	GPT-5 mini	7.3%	estimated ± 6.5 pp, low confidence
133	GPT-5.1-Codex	4.8%	estimated ± 6.5 pp, low confidence
134	Claude Haiku 4.5 Thinking	2.2%	estimated ± 6.5 pp, low confidence
135	o1	0.1%	estimated ± 8.9 pp, low confidence
136	Step 3.7 Flash	0.1%	estimated ± 8.9 pp, low confidence
137	Gemma 4 26B A4B	0.0%	estimated ± 8.9 pp, low confidence
138	DeepSeek V3.2 (Thinking)	0.0%	estimated ± 6.5 pp, low confidence
139	Qwen3 Max	0.0%	estimated ± 6.5 pp, low confidence
140	Nemotron 3 Super 100B	0.0%	estimated ± 8.9 pp, low confidence
141	GLM-4.6	0.0%	estimated ± 6.5 pp, low confidence
142	Celeris-1	0.0%	estimated ± 8.9 pp, low confidence
143	Claude 3 Opus	0.0%	estimated ± 8.9 pp, low confidence
144	Command A+	0.0%	estimated ± 8.9 pp, low confidence
145	Gemini 1.5 Pro	0.0%	estimated ± 8.9 pp, low confidence
146	Gemini 3.1 Flash-Lite	0.0%	estimated ± 6.5 pp, low confidence
147	Gemma 3 27B	0.0%	estimated ± 8.9 pp, low confidence
148	Gemma 4 12B	0.0%	estimated ± 8.9 pp, low confidence
149	Gemma 4 E2B	0.0%	estimated ± 8.9 pp, low confidence
150	Gemma 4 E4B	0.0%	estimated ± 8.9 pp, low confidence
151	GPT-4.1 nano	0.0%	estimated ± 8.9 pp, low confidence
152	GPT-4 Turbo	0.0%	estimated ± 8.9 pp, low confidence
153	GPT-4o mini	0.0%	estimated ± 8.9 pp, low confidence
154	GPT-OSS 120B	0.0%	estimated ± 8.9 pp, low confidence
155	GPT-OSS 20B	0.0%	estimated ± 8.9 pp, low confidence
156	Grok 4.1 Fast (Reasoning)	0.0%	estimated ± 6.5 pp, low confidence
157	Grok 4 Fast (Reasoning)	0.0%	estimated ± 6.5 pp, low confidence
158	K-Exaone	0.0%	estimated ± 8.9 pp, low confidence
159	LFM2.5-2.6B	0.0%	estimated ± 8.9 pp, low confidence
160	Ling 2.6 Flash	0.0%	estimated ± 8.9 pp, low confidence
161	Llama 4 Maverick	0.0%	estimated ± 8.9 pp, low confidence
162	Llama 4 Scout	0.0%	estimated ± 8.9 pp, low confidence
163	Mistral Large 3	0.0%	estimated ± 8.9 pp, low confidence
164	Mistral Small 4	0.0%	estimated ± 8.9 pp, low confidence
165	Mistral Small 4 (Reasoning)	0.0%	estimated ± 8.9 pp, low confidence
166	Nemotron 3 Nano 30B	0.0%	estimated ± 8.9 pp, low confidence
167	Nemotron 3 Nano Omni 30B A3B	0.0%	estimated ± 8.9 pp, low confidence
168	o1-preview	0.0%	estimated ± 8.9 pp, low confidence
169	Trinity-Large-Preview	0.0%	estimated ± 8.9 pp, low confidence
170	Trinity-Large-Thinking	0.0%	estimated ± 8.9 pp, low confidence
171	Ultravox v0.6 Llama 3.3 70B	0.0%	estimated ± 8.9 pp, low confidence

SWE-Rebench leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General