benchgap
Coding

OpenHarmony Bench v1.0 leaderboard

As of 2026-11-18, the highest measured score on OpenHarmony Bench v1.0 is 60.8% by GLM-5.3. 152 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Claude Mythos 560.9%estimated ± 4.9 pp, low confidence
2Sakana Fugu-Ultra60.9%estimated ± 4.9 pp, low confidence
3Beam 1.0 501B-A23B FP860.9%estimated ± 4.9 pp, low confidence
4Beam 1.0 501B-A23B NVFP460.9%estimated ± 4.9 pp, low confidence
5GLM-5.360.8%measured
6Qwen3.8 Max60.8%measured
7DeepSeek V4.1 Flash 101760.3%measured
8ERNIE 5.5 Flash Preview59.8%estimated ± 2.7 pp, high confidence
9LongCat-2.559.3%estimated ± 2.7 pp, high confidence
10Ornith-1.5-397B59.2%estimated ± 2.7 pp, high confidence
11Hy4 preview59.1%estimated ± 2.7 pp, high confidence
12DeepSeek V4 Pro59.0%measured
13Claude Fable 5.258.9%estimated ± 2.4 pp, medium confidence
14Claude Opus 5.558.9%estimated ± 2.4 pp, medium confidence
15Claude Fable 5.158.9%estimated ± 2.4 pp, medium confidence
16Muse Wonder Full58.9%estimated ± 2.4 pp, medium confidence
17Gemini 4 Argon58.9%estimated ± 2.4 pp, medium confidence
18Claude Fable 558.9%estimated ± 2.4 pp, medium confidence
19Claude Sonnet 5.558.9%estimated ± 2.4 pp, medium confidence
20MiMo-V2.6-Pro58.9%estimated ± 2.4 pp, medium confidence
21Kimi K3.1 Pro58.9%estimated ± 2.4 pp, medium confidence
22Muse Wonder58.9%estimated ± 2.4 pp, medium confidence
23Step 5 Preview58.9%estimated ± 2.4 pp, high confidence
24Muse Spark 1.158.9%estimated ± 2.4 pp, high confidence
25Muse Spark 1.358.9%estimated ± 2.4 pp, high confidence
26Gemini 3.1 Pro58.9%estimated ± 2.4 pp, high confidence
27Muse Spark Open58.9%estimated ± 2.4 pp, high confidence
28GPT-6 Sol58.9%estimated ± 2.4 pp, high confidence
29Grok 4.758.9%estimated ± 2.4 pp, high confidence
30Muse Spark 1.258.9%estimated ± 2.4 pp, high confidence
31Gemini 3.7 Flash58.9%estimated ± 2.4 pp, high confidence
32GPT-5.6 Sol58.9%estimated ± 2.4 pp, high confidence
33Gemini 3.8 Flash58.9%estimated ± 2.4 pp, high confidence
34GPT-6 Astra58.9%estimated ± 2.4 pp, high confidence
35Grok 4.658.9%estimated ± 2.4 pp, high confidence
36Claude Opus 558.9%estimated ± 2.4 pp, high confidence
37GPT-5.558.9%estimated ± 2.4 pp, high confidence
38GPT-6.1 Terra58.9%estimated ± 2.4 pp, high confidence
39Claude Haiku 5.558.9%estimated ± 2.4 pp, high confidence
40GPT-5.6 Terra58.9%estimated ± 2.4 pp, high confidence
41Grok 4.558.9%estimated ± 2.4 pp, high confidence
42GPT-6 Luna58.9%estimated ± 2.4 pp, high confidence
43Claude Opus 4.858.9%estimated ± 2.4 pp, high confidence
44Claude Sonnet 558.9%estimated ± 2.4 pp, high confidence
45GPT-6.1 Sol58.9%estimated ± 2.4 pp, high confidence
46Ling 3.1 Flash58.9%estimated ± 2.4 pp, high confidence
47Gemini 3.5 Flash58.9%estimated ± 2.4 pp, high confidence
48GPT-5.6 Luna58.9%estimated ± 2.4 pp, high confidence
49Gemini 3.6 Flash58.9%estimated ± 2.4 pp, high confidence
50MiniMax M3.1 Flash58.9%estimated ± 2.4 pp, high confidence
51Inkling 2 Small58.8%estimated ± 2.4 pp, high confidence
52GPT-5.4 mini58.8%estimated ± 2.4 pp, high confidence
53GLM-5.258.4%measured
54Kimi K2.658.4%estimated ± 2.4 pp, high confidence
55MiMo-V2.6-Flash58.1%estimated ± 2.4 pp, high confidence
56Claude Opus 4.758.1%estimated ± 2.8 pp, high confidence
57GLM-5.3-Flash57.3%measured
58Kimi K357.3%measured
59GPT-5.456.9%estimated ± 2.8 pp, high confidence
60MiMo-V2.5-Pro56.4%estimated ± 2.4 pp, high confidence
61Qwen3.8-Flash-Next56.4%estimated ± 2.4 pp, high confidence
62dots3-note Preview56.3%estimated ± 2.7 pp, high confidence
63Qwen3.8 Max Preview56.0%measured
64Qwen3.8-Omni-Flash55.8%estimated ± 2.7 pp, high confidence
65Ripple-65B Preview55.7%estimated ± 4.9 pp, low confidence
66Ornith-1.0-397B55.4%estimated ± 2.7 pp, high confidence
67Sage-A1.555.0%estimated ± 4.9 pp, low confidence
68Seed 2.1 Pro54.6%estimated ± 2.7 pp, high confidence
69MiniMax M2.754.5%estimated ± 2.4 pp, high confidence
70Composer 2.554.2%estimated ± 3.2 pp, high confidence
71Ornith-1.5-35B-A3B54.0%estimated ± 2.7 pp, high confidence
72DeepSeek V4 Flash 073153.8%measured
73Claude Opus 4.6 (Adaptive)53.7%estimated ± 3.2 pp, high confidence
74GPT-5.3 Codex53.7%estimated ± 3.2 pp, high confidence
75Claude Sonnet 4.653.5%estimated ± 3.2 pp, high confidence
76Qwen3.7 Max53.4%measured
77Atria Dawn Preview53.4%estimated ± 4.9 pp, low confidence
78Laguna S 2.153.3%estimated ± 4.9 pp, low confidence
79Sakana Fugu53.3%estimated ± 4.9 pp, low confidence
80Spark-X2.6-4B53.3%estimated ± 4.9 pp, low confidence
81Claude Opus 4.653.3%estimated ± 4.9 pp, low confidence
82MAI-Thinking-153.3%estimated ± 4.9 pp, low confidence
83Laguna XS 2.153.3%estimated ± 4.9 pp, low confidence
84LongCat-Flash-Lite-Sparse53.3%estimated ± 4.9 pp, low confidence
85Claude Opus 4.5 Thinking53.1%estimated ± 3.2 pp, high confidence
86Gemini 3 Pro53.1%estimated ± 3.2 pp, high confidence
87Inkling-Small53.1%estimated ± 2.4 pp, high confidence
88GPT-5.253.0%estimated ± 3.2 pp, high confidence
89Gemini 3 Flash52.7%estimated ± 3.2 pp, high confidence
90GLM-5.152.3%measured
91Kimi K2.7 Code52.1%measured
92GPT-5.2-Codex51.9%estimated ± 3.2 pp, high confidence
93Grok 4.2051.8%estimated ± 3.2 pp, high confidence
94Seed 2.1 Turbo51.7%estimated ± 2.7 pp, high confidence
95MiMo-V2.551.5%estimated ± 3.2 pp, high confidence
96Muse Spark51.4%estimated ± 2.8 pp, high confidence
97Hy351.3%estimated ± 2.4 pp, high confidence
98Hy3 Preview51.3%estimated ± 2.4 pp, high confidence
99Beam 1.0 501B-A23B51.3%estimated ± 2.4 pp, high confidence
100Claude Opus 4.551.2%estimated ± 2.7 pp, high confidence
101Grok 4.351.2%estimated ± 2.4 pp, high confidence
102Claude Sonnet 4.5 Thinking51.1%estimated ± 3.2 pp, high confidence
103Quasar 438B51.1%estimated ± 2.4 pp, high confidence
104Solar Pro 451.1%estimated ± 2.4 pp, high confidence
105Nemotron 3.5 Super VL 120B A12B51.1%estimated ± 2.4 pp, high confidence
106Nemotron 3.5 Super VL 120B A12B NVFP451.1%estimated ± 2.4 pp, high confidence
107GPT-5.4 nano51.0%estimated ± 2.4 pp, high confidence
108Inkling51.0%estimated ± 2.4 pp, high confidence
109Qwen3.8-27B51.0%estimated ± 2.4 pp, high confidence
110Gemini 2.5 Pro51.0%estimated ± 2.4 pp, high confidence
111Qwen3.7 Plus51.0%estimated ± 2.4 pp, high confidence
112Apodex 1.151.0%estimated ± 2.4 pp, high confidence
113Apodex 1.1 Mini51.0%estimated ± 2.4 pp, high confidence
114Gemma 4 31B51.0%estimated ± 2.4 pp, high confidence
115Muse Glimmer 30B51.0%estimated ± 2.4 pp, high confidence
116Qwen3.5 397B51.0%estimated ± 2.4 pp, high confidence
117Ling 3.0 Flash VL51.0%estimated ± 2.4 pp, medium confidence
118Step 3.7 Flash51.0%estimated ± 2.4 pp, medium confidence
119Qwen3.6-27B51.0%estimated ± 2.4 pp, medium confidence
120K-EXAONE 2.051.0%estimated ± 2.4 pp, medium confidence
121Ling 3.0 Flash51.0%estimated ± 2.4 pp, medium confidence
122Gemini 3.5 Flash-Lite51.0%estimated ± 2.4 pp, medium confidence
123A.X K251.0%estimated ± 2.4 pp, medium confidence
124Trinity-Large-Preview51.0%estimated ± 2.4 pp, medium confidence
125Trinity-Large-Thinking51.0%estimated ± 2.4 pp, medium confidence
126Nemotron 3 Ultra51.0%estimated ± 2.4 pp, medium confidence
127Mistral Medium 3.5 128B51.0%estimated ± 2.4 pp, medium confidence
128Gemma 4 26B A4B51.0%estimated ± 2.4 pp, medium confidence
129Nemotron 3.5 Bolt 74B-A9B51.0%estimated ± 2.4 pp, medium confidence
130Qwen3.5-122B-A10B51.0%estimated ± 2.4 pp, medium confidence
131DeepSeek V3 032451.0%estimated ± 2.4 pp, medium confidence
132GPT-OSS 20B51.0%estimated ± 2.4 pp, medium confidence
133Mistral Small 451.0%estimated ± 2.4 pp, medium confidence
134North Mini Code51.0%estimated ± 2.4 pp, medium confidence
135Command A+51.0%estimated ± 2.4 pp, medium confidence
136Mercury 2.551.0%estimated ± 2.4 pp, medium confidence
137Granite 4.2 30B51.0%estimated ± 2.4 pp, medium confidence
138Mistral Large 351.0%estimated ± 2.4 pp, medium confidence
139Qwen3.6-35B-A3B51.0%estimated ± 2.4 pp, medium confidence
140DeepSeek V351.0%estimated ± 2.4 pp, medium confidence
141GPT-OSS 120B51.0%estimated ± 2.4 pp, medium confidence
142Granite 4.2 8B51.0%estimated ± 2.4 pp, medium confidence
143Llama 4 Maverick51.0%estimated ± 2.4 pp, medium confidence
144Nemotron 3.5 Lightning 30B A3B NVFP451.0%estimated ± 2.4 pp, medium confidence
145Qwen 3.6 Max (preview)50.8%estimated ± 2.7 pp, high confidence
146Claude Haiku 4.550.1%estimated ± 3.2 pp, medium confidence
147Qwen3.5 Flash49.4%estimated ± 3.2 pp, medium confidence
148Gemini 3.1 Flash-Lite48.9%estimated ± 3.2 pp, medium confidence
149MiniMax M348.4%measured
150MiMo-V2-Flash47.4%estimated ± 2.8 pp, medium confidence
151Grok 447.3%estimated ± 3.2 pp, medium confidence
152Laguna M.147.3%estimated ± 3.2 pp, medium confidence
153GPT-5.147.2%estimated ± 2.8 pp, medium confidence
154Laguna XS.246.5%estimated ± 3.2 pp, medium confidence
155GLM-4.745.3%estimated ± 2.8 pp, medium confidence
156Grok 4 Fast (Reasoning)43.4%estimated ± 3.2 pp, medium confidence
157Qwen3.6 Plus43.2%estimated ± 2.7 pp, medium confidence
158o142.8%estimated ± 2.8 pp, medium confidence
159Grok 4.1 Fast (Reasoning)42.2%estimated ± 3.2 pp, medium confidence
160GPT-542.0%estimated ± 2.8 pp, medium confidence
161GLM-539.0%estimated ± 2.7 pp, medium confidence
162Ornith-1.0-35B35.9%estimated ± 2.7 pp, medium confidence
163Ornith-1.5-9B30.1%estimated ± 2.7 pp, medium confidence
164Kimi K2.529.0%estimated ± 2.7 pp, medium confidence
165Ornith-1.0-9B15.8%estimated ± 2.7 pp, medium confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General