benchgap
Agentic · tools

DRACO leaderboard

As of 2026-11-21, the highest measured score on DRACO is 88.6% by Claude Opus 5. 167 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Claude Opus 588.6%measured
2Atria Dawn Preview88.0%estimated ± 2.9 pp, low confidence
3GPT-6 Astra87.9%estimated ± 0.7 pp, low confidence
4GPT-6.2 Sol87.9%estimated ± 0.7 pp, low confidence
5Gemini 4 Neon87.9%estimated ± 0.7 pp, low confidence
6GPT-6.1 Sol87.9%estimated ± 0.7 pp, low confidence
7GPT-6.2 Astra87.9%estimated ± 0.7 pp, low confidence
8Muse Wonder Full87.9%estimated ± 0.7 pp, low confidence
9Gemini 4 Carbon87.9%estimated ± 0.7 pp, low confidence
10GPT-6.1 Terra87.9%estimated ± 0.7 pp, low confidence
11Gemini 4 Argon87.9%estimated ± 0.7 pp, low confidence
12GPT-5.6 Sol87.9%estimated ± 0.7 pp, low confidence
13Claude Fable 587.9%estimated ± 0.7 pp, low confidence
14Claude Fable 5.187.7%measured
15Claude Fable 5.287.7%measured
16LongCat-2.587.5%estimated ± 2.9 pp, low confidence
17Claude Opus 5.587.4%measured
18Claude Sonnet 5.587.0%measured
19GPT-6 Sol86.9%estimated ± 0.7 pp, low confidence
20GPT-5.6 Terra86.9%estimated ± 0.7 pp, low confidence
21Gemini 3.8 Flash86.9%estimated ± 0.7 pp, low confidence
22Gemini 4 Hydrogen86.9%estimated ± 0.7 pp, low confidence
23Kimi K3.1 Pro86.9%estimated ± 0.7 pp, low confidence
24Claude Haiku 5.586.9%estimated ± 0.7 pp, low confidence
25DeepSeek V4.1 Flash 101786.9%estimated ± 0.7 pp, low confidence
26Gemini 3.1 Pro86.9%estimated ± 0.7 pp, low confidence
27Gemini 3.7 Flash86.9%estimated ± 0.7 pp, low confidence
28Grok 4.686.9%estimated ± 0.7 pp, low confidence
29Grok 4.786.9%estimated ± 0.7 pp, low confidence
30Kimi K386.9%estimated ± 0.7 pp, low confidence
31Claude Haiku 4.586.9%estimated ± 0.7 pp, low confidence
32Claude Opus 4.886.9%estimated ± 0.7 pp, low confidence
33Claude Sonnet 4.686.9%estimated ± 0.7 pp, low confidence
34Claude Sonnet 586.9%estimated ± 0.7 pp, low confidence
35DeepSeek V4 Flash 073186.9%estimated ± 0.7 pp, low confidence
36Gemini 3.5 Flash86.9%estimated ± 0.7 pp, low confidence
37Gemini 3.6 Flash86.9%estimated ± 0.7 pp, low confidence
38GLM-5.286.9%estimated ± 0.7 pp, low confidence
39GLM-5.386.9%estimated ± 0.7 pp, low confidence
40GLM-5.3-Flash86.9%estimated ± 0.7 pp, low confidence
41GPT-5.3-Codex-Spark86.9%estimated ± 0.7 pp, low confidence
42GPT-5.486.9%estimated ± 0.7 pp, low confidence
43GPT-5.586.9%estimated ± 0.7 pp, low confidence
44GPT-5.6 Luna86.9%estimated ± 0.7 pp, low confidence
45GPT-6 Luna86.9%estimated ± 0.7 pp, low confidence
46Grok 4.2086.9%estimated ± 0.7 pp, low confidence
47Llama 4 Maverick86.9%estimated ± 0.7 pp, low confidence
48MiMo-V2.5-Pro86.9%estimated ± 0.7 pp, low confidence
49MiniMax M386.9%estimated ± 0.7 pp, low confidence
50Muse Ember86.9%estimated ± 0.7 pp, low confidence
51Muse Spark 1.286.9%estimated ± 0.7 pp, low confidence
52Muse Spark 1.386.9%estimated ± 0.7 pp, low confidence
53Muse Spark Open86.9%estimated ± 0.7 pp, low confidence
54Muse Wonder86.9%estimated ± 0.7 pp, low confidence
55Nemotron 3 Ultra86.9%estimated ± 0.7 pp, low confidence
56Qwen3.8-27B86.9%estimated ± 0.7 pp, low confidence
57Qwen3.8 Max86.9%estimated ± 0.7 pp, low confidence
58MiMo-V2.6-Pro86.2%estimated ± 2.0 pp, medium confidence
59Qwen3.8 Max Preview86.0%estimated ± 2.0 pp, medium confidence
60Inkling 2 Small86.0%estimated ± 2.9 pp, medium confidence
61GPT-5.5 Pro85.9%estimated ± 2.9 pp, medium confidence
62Qwen3.8-Flash-Next85.6%estimated ± 2.0 pp, medium confidence
63Ling 3.1 Flash85.5%measured
64MiMo-V2.6-Flash85.4%estimated ± 2.0 pp, medium confidence
65GPT-5.4 Pro85.2%estimated ± 2.9 pp, medium confidence
66DeepSeek V4 Pro85.2%estimated ± 2.0 pp, medium confidence
67ERNIE 5.5 Flash Preview84.9%measured
68Claude Mythos 584.1%estimated ± 2.9 pp, medium confidence
69MiniMax M3.1 Flash83.5%estimated ± 2.0 pp, medium confidence
70Step 5 Preview83.3%measured
71Mistral Large 483.0%estimated ± 2.0 pp, medium confidence
72Ornith-1.5-397B82.9%estimated ± 2.9 pp, medium confidence
73Grok 4.582.4%estimated ± 2.0 pp, medium confidence
74Claude Opus 4.781.8%estimated ± 2.0 pp, medium confidence
75Claude Opus 4.680.4%estimated ± 2.9 pp, medium confidence
76dots3-note Preview80.1%estimated ± 2.9 pp, medium confidence
77Hy3 Preview79.1%estimated ± 2.0 pp, medium confidence
78Muse Spark 1.179.0%estimated ± 2.0 pp, medium confidence
79Sage-A1.578.7%estimated ± 2.9 pp, medium confidence
80Apodex 1.178.4%estimated ± 2.0 pp, medium confidence
81Apodex 1.1 Mini78.4%estimated ± 2.0 pp, medium confidence
82Quasar 438B78.3%estimated ± 2.0 pp, medium confidence
83Beam 1.0 501B-A23B78.2%estimated ± 2.0 pp, medium confidence
84Nemotron 3.5 Super VL 120B A12B77.8%estimated ± 2.0 pp, medium confidence
85Nemotron 3.5 Super VL 120B A12B NVFP477.8%estimated ± 2.0 pp, medium confidence
86Ling 3.0 Flash VL77.7%estimated ± 2.0 pp, medium confidence
87Ling 3.0 Flash Sante77.4%estimated ± 2.9 pp, medium confidence
88Hy4 preview77.2%measured
89Qwen3.7 Max76.8%estimated ± 2.0 pp, medium confidence
90Inkling-Small76.7%estimated ± 2.0 pp, medium confidence
91GLM-5.176.5%estimated ± 2.0 pp, medium confidence
92Inkling 2 Mini76.4%estimated ± 2.9 pp, medium confidence
93Solar Pro 476.2%estimated ± 2.0 pp, medium confidence
94Ripple-65B Preview75.8%estimated ± 4.0 pp, medium confidence
95Grok 4.375.5%estimated ± 2.0 pp, medium confidence
96Inkling75.1%estimated ± 2.0 pp, medium confidence
97Hy374.9%estimated ± 2.0 pp, medium confidence
98Beam 1.0 501B-A23B FP874.4%estimated ± 2.9 pp, medium confidence
99Kimi K2.674.0%estimated ± 2.0 pp, medium confidence
100Kimi K2.7 Code74.0%estimated ± 2.0 pp, medium confidence
101Beam 1.0 501B-A23B NVFP473.5%estimated ± 2.9 pp, medium confidence
102Agents-A173.4%estimated ± 2.9 pp, medium confidence
103Step 3.7 Flash73.1%estimated ± 2.0 pp, medium confidence
104GLM-4.773.1%estimated ± 2.0 pp, medium confidence
105GPT-5.4 mini73.0%estimated ± 2.0 pp, medium confidence
106MiniMax M2.772.9%estimated ± 2.0 pp, medium confidence
107Muse Spark72.4%estimated ± 2.0 pp, medium confidence
108Qwen3.6 Plus72.1%estimated ± 2.0 pp, medium confidence
109Qwen3.6-27B71.9%estimated ± 2.0 pp, medium confidence
110Gemini 3.5 Flash-Lite71.9%estimated ± 2.0 pp, medium confidence
111Nemotron 3.5 Bolt 74B-A9B71.7%estimated ± 2.0 pp, medium confidence
112A.X K270.8%estimated ± 2.0 pp, medium confidence
113Ling 3.0 Flash70.4%measured
114GPT-5.4 nano70.3%estimated ± 2.0 pp, medium confidence
115GPT-568.8%estimated ± 2.0 pp, low confidence
116Nemotron 3.5 Bolt 74B-A9B BF1668.7%estimated ± 2.9 pp, low confidence
117Qwen3.5 397B67.8%estimated ± 2.9 pp, low confidence
118Qwen3.6-35B-A3B67.4%estimated ± 2.0 pp, low confidence
119Ornith-1.5-35B-A3B66.6%estimated ± 2.9 pp, low confidence
120Solar Open 266.2%estimated ± 4.0 pp, low confidence
121Agents-A1-4B65.9%estimated ± 2.9 pp, low confidence
122GPT-5.265.1%estimated ± 2.9 pp, low confidence
123Kimi K2.564.2%estimated ± 2.0 pp, low confidence
124GPT-5.163.2%estimated ± 2.0 pp, low confidence
125Qwen3.5-122B-A10B62.4%estimated ± 2.0 pp, low confidence
126K-EXAONE 2.061.9%estimated ± 2.0 pp, low confidence
127GLM-561.8%estimated ± 2.9 pp, low confidence
128Qwen3.5-27B60.9%estimated ± 2.9 pp, low confidence
129Qwen3.5-35B-A3B60.9%estimated ± 2.9 pp, low confidence
130Muse Glimmer 30B60.0%estimated ± 2.0 pp, low confidence
131Qwen3.7 Plus58.4%estimated ± 2.0 pp, low confidence
132Mistral Medium 3.5 128B57.8%estimated ± 2.0 pp, low confidence
133Ornith-1.5-9B57.0%estimated ± 2.9 pp, low confidence
134LLaDA2.2-flash56.4%estimated ± 4.0 pp, low confidence
135LongCat-Flash-Lite-Sparse55.9%estimated ± 4.0 pp, low confidence
136Spark-X2.6-4B54.2%estimated ± 2.9 pp, low confidence
137MiniCPM5-2B52.3%estimated ± 2.0 pp, low confidence
138Nemotron 3.5 Lightning 30B A3B NVFP442.1%estimated ± 2.0 pp, low confidence
139MiMo-V2-Flash40.3%estimated ± 2.0 pp, low confidence
140Gemma 4 31B38.4%estimated ± 2.0 pp, low confidence
141GPT-OSS 120B36.3%estimated ± 2.0 pp, low confidence
142Granite 4.2 30B28.1%estimated ± 2.0 pp, low confidence
143Gemma 4 26B A4B25.2%estimated ± 2.0 pp, low confidence
144Ling 3.0 Tiny24.0%estimated ± 2.0 pp, low confidence
145Command A+6.9%estimated ± 2.0 pp, low confidence
146Celeris-10.0%estimated ± 2.0 pp, low confidence
147Claude 3.5 Haiku0.0%estimated ± 2.0 pp, low confidence
148DeepSeek V30.0%estimated ± 2.0 pp, low confidence
149DeepSeek V3 03240.0%estimated ± 2.0 pp, low confidence
150Gemini 2.5 Pro0.0%estimated ± 2.0 pp, low confidence
151Gemma 3 27B0.0%estimated ± 2.0 pp, low confidence
152Gemma 4 12B0.0%estimated ± 2.0 pp, low confidence
153Gemma 4 E2B0.0%estimated ± 2.0 pp, low confidence
154Gemma 4 E4B0.0%estimated ± 2.0 pp, low confidence
155GPT-4.1 mini0.0%estimated ± 2.0 pp, low confidence
156GPT-4.1 nano0.0%estimated ± 2.0 pp, low confidence
157GPT-4o0.0%estimated ± 2.0 pp, low confidence
158GPT-4o mini0.0%estimated ± 2.0 pp, low confidence
159GPT-OSS 20B0.0%estimated ± 2.0 pp, low confidence
160Granite 4.2 3B0.0%estimated ± 2.0 pp, low confidence
161Granite 4.2 8B0.0%estimated ± 2.0 pp, low confidence
162K-Exaone0.0%estimated ± 2.0 pp, low confidence
163LFM2.5-2.6B0.0%estimated ± 2.0 pp, low confidence
164Ling 2.6 Flash0.0%estimated ± 2.0 pp, low confidence
165Llama 3.3 70B0.0%estimated ± 2.0 pp, low confidence
166Llama 4 Scout0.0%estimated ± 2.0 pp, low confidence
167Mercury 2.50.0%estimated ± 2.0 pp, low confidence
168Mistral Large 30.0%estimated ± 2.0 pp, low confidence
169Mistral Small 3.10.0%estimated ± 2.0 pp, low confidence
170Mistral Small 40.0%estimated ± 2.0 pp, low confidence
171Nemotron 3 Nano 30B0.0%estimated ± 2.0 pp, low confidence
172Nemotron 3 Nano Omni 30B A3B0.0%estimated ± 2.0 pp, low confidence
173North Mini Code0.0%estimated ± 2.0 pp, low confidence
174Solar Pro 30.0%estimated ± 2.0 pp, low confidence
175Trinity-Large-Preview0.0%estimated ± 2.0 pp, low confidence
176Trinity-Large-Thinking0.0%estimated ± 2.0 pp, low confidence
177Ultravox v0.6 Llama 3.3 70B0.0%estimated ± 2.0 pp, low confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General