Coding

OpenHarmony Bench leaderboard

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

As of 2026-10-07, the highest measured score on OpenHarmony Bench is 60.8% by GLM-5.3. 157 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Mythos 5	60.9%	estimated ± 4.9 pp, low confidence
2	Sakana Fugu-Ultra	60.9%	estimated ± 4.9 pp, low confidence
3	Beam	60.9%	estimated ± 4.9 pp, low confidence
4	GLM-5.3	60.8%	measured
5	Qwen3.8 Max	60.8%	measured
6	DeepSeek V4.1 Flash	60.3%	measured
7	Ornith-1.5-397B	59.3%	estimated ± 2.7 pp, high confidence
8	Hy4 preview	59.1%	estimated ± 2.7 pp, high confidence
9	DeepSeek V4 Pro 0813	59.0%	measured
10	Claude Fable 5	58.8%	estimated ± 2.4 pp, medium confidence
11	Claude Fable 5.1	58.8%	estimated ± 2.4 pp, medium confidence
12	Claude Opus 5	58.8%	estimated ± 2.4 pp, high confidence
13	Claude Opus 5.5	58.8%	estimated ± 2.4 pp, medium confidence
14	Claude Sonnet 5.5	58.8%	estimated ± 2.4 pp, medium confidence
15	Gemini 3.1 Pro	58.8%	estimated ± 2.4 pp, high confidence
16	Gemini 3.7 Flash	58.8%	estimated ± 2.4 pp, high confidence
17	Gemini 3.8 Flash	58.8%	estimated ± 2.4 pp, high confidence
18	Gemini 4 Argon	58.8%	estimated ± 2.4 pp, medium confidence
19	GPT-5.6 Sol	58.8%	estimated ± 2.4 pp, high confidence
20	GPT-6 Astra	58.8%	estimated ± 2.4 pp, high confidence
21	GPT-6 Sol	58.8%	estimated ± 2.4 pp, high confidence
22	Grok 4.6	58.8%	estimated ± 2.4 pp, high confidence
23	Grok 4.7	58.8%	estimated ± 2.4 pp, high confidence
24	MiMo-V2.6-Pro	58.8%	estimated ± 2.4 pp, medium confidence
25	Muse Spark 1.1	58.8%	estimated ± 2.4 pp, high confidence
26	Muse Spark 1.2	58.8%	estimated ± 2.4 pp, high confidence
27	Muse Spark 1.3	58.8%	estimated ± 2.4 pp, high confidence
28	Step 5 Preview	58.8%	estimated ± 2.4 pp, high confidence
29	GPT-5.5	58.8%	estimated ± 2.4 pp, high confidence
30	Claude Haiku 5.5	58.8%	estimated ± 2.4 pp, high confidence
31	GPT-5.6 Terra	58.8%	estimated ± 2.4 pp, high confidence
32	Grok 4.5	58.8%	estimated ± 2.4 pp, high confidence
33	GPT-6 Luna	58.8%	estimated ± 2.4 pp, high confidence
34	Claude Opus 4.8	58.8%	estimated ± 2.4 pp, high confidence
35	Claude Sonnet 5	58.8%	estimated ± 2.4 pp, high confidence
36	GPT-6.1 Sol	58.8%	estimated ± 2.4 pp, high confidence
37	Mistral Large 4	58.8%	estimated ± 2.4 pp, high confidence
38	Ling 3.1 Flash	58.8%	estimated ± 2.4 pp, high confidence
39	Gemini 3.5 Flash	58.8%	estimated ± 2.4 pp, high confidence
40	GPT-5.6 Luna	58.7%	estimated ± 2.4 pp, high confidence
41	Gemini 3.6 Flash	58.7%	estimated ± 2.4 pp, high confidence
42	GPT-5.4 mini	58.6%	estimated ± 2.4 pp, high confidence
43	GLM-5.2	58.4%	measured
44	Kimi K2.6	58.2%	estimated ± 2.4 pp, high confidence
45	Claude Opus 4.7 (Adaptive)	58.1%	estimated ± 2.8 pp, high confidence
46	MiMo-V2.6-Flash	58.0%	estimated ± 2.4 pp, high confidence
47	GLM-5.3-Flash	57.3%	measured
48	Kimi K3	57.3%	measured
49	GPT-5.4	56.9%	estimated ± 2.8 pp, high confidence
50	MiMo-V2.5-Pro	56.4%	estimated ± 2.4 pp, high confidence
51	Qwen3.8-Flash-Next	56.4%	estimated ± 2.4 pp, high confidence
52	Qwen3.8 Max Preview	56.0%	measured
53	dots3-note Preview	55.2%	estimated ± 2.7 pp, high confidence
54	Claude Opus 4.7	54.9%	estimated ± 3.2 pp, high confidence
55	Qwen3.8-Omni-Flash	54.7%	estimated ± 2.7 pp, high confidence
56	MiniMax M2.7	54.5%	estimated ± 2.4 pp, high confidence
57	Ornith-1.0-397B	54.4%	estimated ± 2.7 pp, high confidence
58	Composer 2.5	54.2%	estimated ± 3.2 pp, high confidence
59	DeepSeek V4 Flash 0731	53.8%	measured
60	Seed 2.1 Pro	53.8%	estimated ± 2.7 pp, high confidence
61	GPT-5.3 Codex	53.7%	estimated ± 3.2 pp, high confidence
62	Claude Sonnet 4.6	53.5%	estimated ± 3.2 pp, high confidence
63	Qwen3.7 Max	53.4%	measured
64	Ornith-1.5-35B-A3B	53.4%	estimated ± 2.7 pp, high confidence
65	Atria Dawn Preview	53.4%	estimated ± 4.9 pp, low confidence
66	Laguna S 2.1	53.3%	estimated ± 4.9 pp, low confidence
67	Sakana Fugu	53.3%	estimated ± 4.9 pp, low confidence
68	Claude Opus 4.6	53.3%	estimated ± 4.9 pp, low confidence
69	GLM-5	53.3%	estimated ± 4.9 pp, low confidence
70	GPT-5.2	53.3%	estimated ± 4.9 pp, low confidence
71	Laguna XS 2.1	53.3%	estimated ± 4.9 pp, low confidence
72	LLaDA2.2-flash	53.3%	estimated ± 4.9 pp, low confidence
73	LongCat-Flash-Lite-Sparse	53.3%	estimated ± 4.9 pp, low confidence
74	MAI-Thinking-1	53.3%	estimated ± 4.9 pp, low confidence
75	Qwen3.5 397B	53.3%	estimated ± 4.9 pp, low confidence
76	Inkling-Small	53.1%	estimated ± 2.4 pp, high confidence
77	Gemini 3 Flash	52.7%	estimated ± 3.2 pp, high confidence
78	GLM-5.1	52.3%	measured
79	Kimi K2.7 Code	52.1%	measured
80	Seed 2.1 Turbo	52.0%	estimated ± 2.7 pp, high confidence
81	GPT-5.2-Codex	51.9%	estimated ± 3.2 pp, high confidence
82	Grok 4.20	51.8%	estimated ± 3.2 pp, high confidence
83	Claude Opus 4.5	51.8%	estimated ± 2.7 pp, high confidence
84	Qwen 3.6 Max (preview)	51.6%	estimated ± 2.7 pp, high confidence
85	MiMo-V2.5	51.5%	estimated ± 3.2 pp, high confidence
86	Muse Spark	51.4%	estimated ± 2.8 pp, high confidence
87	Hy3	51.3%	estimated ± 2.4 pp, high confidence
88	Hy3 Preview	51.3%	estimated ± 2.4 pp, high confidence
89	Grok 4.3	51.2%	estimated ± 2.4 pp, high confidence
90	Quasar 438B	51.1%	estimated ± 2.4 pp, high confidence
91	GPT-5.4 nano	51.0%	estimated ± 2.4 pp, high confidence
92	Inkling	51.0%	estimated ± 2.4 pp, high confidence
93	Qwen3.8-27B	51.0%	estimated ± 2.4 pp, high confidence
94	Gemini 2.5 Pro	51.0%	estimated ± 2.4 pp, high confidence
95	Qwen3.7 Plus	51.0%	estimated ± 2.4 pp, high confidence
96	Apodex 1.1	51.0%	estimated ± 2.4 pp, high confidence
97	Apodex 1.1 Mini	51.0%	estimated ± 2.4 pp, high confidence
98	Gemma 4 31B	51.0%	estimated ± 2.4 pp, high confidence
99	Muse Glimmer 30B	51.0%	estimated ± 2.4 pp, high confidence
100	Solar Pro 4	51.0%	estimated ± 2.4 pp, medium confidence
101	A.X K2	51.0%	estimated ± 2.4 pp, medium confidence
102	Celeris-1	51.0%	estimated ± 2.4 pp, medium confidence
103	Command A+	51.0%	estimated ± 2.4 pp, medium confidence
104	DeepSeek V3	51.0%	estimated ± 2.4 pp, medium confidence
105	DeepSeek V3 0324	51.0%	estimated ± 2.4 pp, medium confidence
106	Gemini 3.5 Flash-Lite	51.0%	estimated ± 2.4 pp, medium confidence
107	Gemma 3 27B	51.0%	estimated ± 2.4 pp, medium confidence
108	Gemma 4 26B A4B	51.0%	estimated ± 2.4 pp, medium confidence
109	Gemma 4 E4B	51.0%	estimated ± 2.4 pp, medium confidence
110	GPT-OSS 120B	51.0%	estimated ± 2.4 pp, medium confidence
111	GPT-OSS 20B	51.0%	estimated ± 2.4 pp, medium confidence
112	Granite 4.2 30B	51.0%	estimated ± 2.4 pp, medium confidence
113	Granite 4.2 3B	51.0%	estimated ± 2.4 pp, medium confidence
114	Granite 4.2 8B	51.0%	estimated ± 2.4 pp, medium confidence
115	K-EXAONE 2.0	51.0%	estimated ± 2.4 pp, medium confidence
116	LFM2.5-2.6B	51.0%	estimated ± 2.4 pp, medium confidence
117	Ling 3.0 Flash	51.0%	estimated ± 2.4 pp, medium confidence
118	Ling 3.0 Flash FP8	51.0%	estimated ± 2.4 pp, medium confidence
119	Ling 3.0 Flash VL	51.0%	estimated ± 2.4 pp, medium confidence
120	Ling 3.0 Tiny	51.0%	estimated ± 2.4 pp, medium confidence
121	Llama 4 Maverick	51.0%	estimated ± 2.4 pp, medium confidence
122	Llama 4 Scout	51.0%	estimated ± 2.4 pp, medium confidence
123	Mercury 2.5	51.0%	estimated ± 2.4 pp, medium confidence
124	MiniCPM5-2B	51.0%	estimated ± 2.4 pp, medium confidence
125	Mistral Large 3	51.0%	estimated ± 2.4 pp, medium confidence
126	Mistral Medium 3.5 128B	51.0%	estimated ± 2.4 pp, medium confidence
127	Mistral Small 4	51.0%	estimated ± 2.4 pp, medium confidence
128	Mistral Small 4 (Reasoning)	51.0%	estimated ± 2.4 pp, medium confidence
129	Nemotron 3.5 Lightning 30B A3B NVFP4	51.0%	estimated ± 2.4 pp, medium confidence
130	Nemotron 3 Nano 30B	51.0%	estimated ± 2.4 pp, medium confidence
131	Nemotron 3 Super 100B	51.0%	estimated ± 2.4 pp, medium confidence
132	Nemotron 3 Ultra	51.0%	estimated ± 2.4 pp, medium confidence
133	North Mini Code	51.0%	estimated ± 2.4 pp, medium confidence
134	Qwen3.5-122B-A10B	51.0%	estimated ± 2.4 pp, medium confidence
135	Qwen3.6-27B	51.0%	estimated ± 2.4 pp, medium confidence
136	Qwen3.6-35B-A3B	51.0%	estimated ± 2.4 pp, medium confidence
137	Solar Pro 3	51.0%	estimated ± 2.4 pp, medium confidence
138	Step 3.7 Flash	51.0%	estimated ± 2.4 pp, medium confidence
139	Trinity-Large-Preview	51.0%	estimated ± 2.4 pp, medium confidence
140	Trinity-Large-Thinking	51.0%	estimated ± 2.4 pp, medium confidence
141	Claude Haiku 4.5	50.1%	estimated ± 3.2 pp, medium confidence
142	Qwen3.6 Plus	49.5%	estimated ± 2.8 pp, medium confidence
143	Qwen3.5 Flash	49.4%	estimated ± 3.2 pp, medium confidence
144	Gemini 3.1 Flash-Lite	48.9%	estimated ± 3.2 pp, medium confidence
145	MiniMax M3	48.4%	measured
146	MiMo-V2-Flash	47.4%	estimated ± 2.8 pp, medium confidence
147	Laguna M.1	47.3%	estimated ± 3.2 pp, medium confidence
148	GPT-5.1	47.2%	estimated ± 2.8 pp, medium confidence
149	Laguna XS.2	46.5%	estimated ± 3.2 pp, medium confidence
150	Ornith-1.0-35B	46.4%	estimated ± 2.7 pp, medium confidence
151	Kimi K2.5	46.0%	estimated ± 2.8 pp, medium confidence
152	Kimi K2.5 (Reasoning)	46.0%	estimated ± 2.8 pp, medium confidence
153	GLM-4.7	45.3%	estimated ± 2.8 pp, medium confidence
154	Ornith-1.5-9B	44.8%	estimated ± 2.7 pp, medium confidence
155	o1	42.8%	estimated ± 2.8 pp, medium confidence
156	GPT-5 (high)	42.0%	estimated ± 2.8 pp, medium confidence
157	Ornith-1.0-9B	40.7%	estimated ± 2.7 pp, medium confidence
158	o1-preview	40.3%	estimated ± 2.8 pp, medium confidence
159	K-Exaone	39.4%	estimated ± 2.8 pp, medium confidence
160	Gemma 4 12B	38.9%	estimated ± 2.8 pp, medium confidence
161	Ling 2.6 Flash	36.3%	estimated ± 2.8 pp, medium confidence
162	Gemini 1.5 Pro	35.6%	estimated ± 2.8 pp, medium confidence
163	GPT-4 Turbo	34.6%	estimated ± 2.8 pp, medium confidence
164	GPT-4.1 mini	34.1%	estimated ± 2.8 pp, medium confidence
165	Claude 3 Opus	33.7%	estimated ± 2.8 pp, medium confidence
166	Nemotron 3 Nano Omni 30B A3B	31.1%	estimated ± 2.8 pp, medium confidence
167	Ultravox v0.6 Llama 3.3 70B	30.3%	estimated ± 2.8 pp, medium confidence
168	GPT-4o mini	30.1%	estimated ± 2.8 pp, medium confidence
169	GPT-4.1 nano	30.0%	estimated ± 2.8 pp, medium confidence
170	Gemma 4 E2B	28.2%	estimated ± 2.8 pp, medium confidence

OpenHarmony Bench leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General