benchgap
General

OTelBench leaderboard

As of 2026-11-18, the highest measured score on OTelBench is 29.0% by Claude Opus 4.5. 99 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Claude Haiku 5.532.1%estimated ± 4.1 pp, low confidence
2Gemini 3.6 Flash32.1%estimated ± 4.1 pp, low confidence
3Grok 4.731.2%estimated ± 4.1 pp, low confidence
4o331.2%estimated ± 4.1 pp, low confidence
5Claude Opus 4.730.3%estimated ± 4.1 pp, low confidence
6Claude Opus 4.529.0%measured
7Kimi K328.4%estimated ± 4.1 pp, low confidence
8Claude Opus 4.627.5%estimated ± 4.1 pp, low confidence
9Muse Spark 1.327.5%estimated ± 4.1 pp, low confidence
10GPT-5.226.1%measured
11GPT-525.6%estimated ± 4.1 pp, medium confidence
12Qwen3.6-35B-A3B24.6%estimated ± 4.1 pp, medium confidence
13GPT-5.6 Luna23.6%estimated ± 4.1 pp, medium confidence
14Nemotron 3 Ultra22.6%estimated ± 4.1 pp, medium confidence
15Qwen3.5 Flash22.6%estimated ± 4.1 pp, medium confidence
16Claude Fable 522.5%estimated ± 6.6 pp, medium confidence
17Claude Fable 5.122.5%estimated ± 6.6 pp, medium confidence
18Claude Fable 5.222.5%estimated ± 6.6 pp, low confidence
19Claude Opus 4.6 (Adaptive)22.5%estimated ± 6.6 pp, medium confidence
20Claude Opus 4.822.5%estimated ± 6.6 pp, medium confidence
21Claude Opus 522.5%estimated ± 6.6 pp, medium confidence
22Claude Opus 5.522.5%estimated ± 6.6 pp, medium confidence
23Claude Sonnet 522.5%estimated ± 6.6 pp, medium confidence
24Claude Sonnet 5.522.5%estimated ± 6.6 pp, medium confidence
25DeepSeek V4.1 Flash 101722.5%estimated ± 6.6 pp, medium confidence
26Gemini 3.5 Flash22.5%estimated ± 6.6 pp, medium confidence
27Gemini 3.7 Flash22.5%estimated ± 6.6 pp, medium confidence
28Gemini 3.8 Flash22.5%estimated ± 6.6 pp, low confidence
29Gemini 4 Argon22.5%estimated ± 6.6 pp, low confidence
30Gemini 4 Carbon22.5%estimated ± 6.6 pp, low confidence
31Gemma 4 31B22.5%estimated ± 6.6 pp, medium confidence
32GLM-5.122.5%estimated ± 6.6 pp, medium confidence
33GLM-5.322.5%estimated ± 6.6 pp, medium confidence
34GPT-5.522.5%estimated ± 6.6 pp, low confidence
35GPT-5.6 Sol22.5%estimated ± 6.6 pp, medium confidence
36GPT-5.6 Terra22.5%estimated ± 6.6 pp, medium confidence
37GPT-6.1 Sol22.5%estimated ± 6.6 pp, low confidence
38GPT-6.1 Terra22.5%estimated ± 6.6 pp, medium confidence
39GPT-6 Astra22.5%estimated ± 6.6 pp, low confidence
40GPT-6 Luna22.5%estimated ± 6.6 pp, medium confidence
41GPT-6 Sol22.5%estimated ± 6.6 pp, medium confidence
42Grok 4.1 Fast (Reasoning)22.5%estimated ± 6.6 pp, medium confidence
43Grok 4.2022.5%estimated ± 6.6 pp, medium confidence
44Grok 4.20 Multi-agent22.5%estimated ± 6.6 pp, medium confidence
45Grok 4.522.5%estimated ± 6.6 pp, medium confidence
46Grok 4.622.5%estimated ± 6.6 pp, medium confidence
47Hy422.5%estimated ± 6.6 pp, medium confidence
48Hy4 preview22.5%estimated ± 6.6 pp, medium confidence
49Kimi K3.1 Pro22.5%estimated ± 6.6 pp, medium confidence
50Muse Spark 1.122.5%estimated ± 6.6 pp, medium confidence
51Muse Spark 1.222.5%estimated ± 6.6 pp, medium confidence
52Muse Spark Open22.5%estimated ± 6.6 pp, medium confidence
53Muse Wonder22.5%estimated ± 6.6 pp, medium confidence
54Qwen3.7 Max22.5%estimated ± 6.6 pp, medium confidence
55Qwen3.8 Max22.5%estimated ± 6.6 pp, medium confidence
56Step 5 Preview22.5%estimated ± 6.6 pp, medium confidence
57MiniMax M322.5%estimated ± 6.6 pp, medium confidence
58GPT-5.4 mini22.5%estimated ± 6.6 pp, medium confidence
59Grok 4.322.5%estimated ± 6.6 pp, medium confidence
60Qwen3.8-27B22.5%estimated ± 6.6 pp, medium confidence
61Claude Opus 4.5 Thinking22.5%estimated ± 6.6 pp, medium confidence
62Qwen3.5-27B22.5%estimated ± 6.6 pp, medium confidence
63Hy322.4%estimated ± 6.6 pp, medium confidence
64Step 3.7 Flash22.4%estimated ± 6.6 pp, medium confidence
65Gemini 2.5 Flash22.4%estimated ± 7.7 pp, low confidence
66Claude Sonnet 4.5 Thinking22.3%estimated ± 6.6 pp, medium confidence
67o4-mini22.2%estimated ± 7.7 pp, low confidence
68Claude Sonnet 4.521.7%measured
69Gemini 3.5 Flash-Lite21.6%estimated ± 4.1 pp, medium confidence
70Qwen 3.6 Max (preview)21.6%estimated ± 4.1 pp, medium confidence
71Grok 4 Fast (Reasoning)21.4%estimated ± 7.7 pp, low confidence
72Kimi K2.620.6%estimated ± 4.1 pp, medium confidence
73Qwen3.5 397B20.6%estimated ± 4.1 pp, medium confidence
74Gemini 3.1 Pro20.1%estimated ± 5.6 pp, low confidence
75GPT-5.3 Codex20.1%estimated ± 5.6 pp, medium confidence
76Kimi K2.519.9%estimated ± 5.6 pp, medium confidence
77Gemini 2.5 Pro19.6%estimated ± 5.6 pp, medium confidence
78MiMo-V2.5-Pro19.6%estimated ± 6.6 pp, medium confidence
79Mistral Large 419.6%estimated ± 4.1 pp, medium confidence
80Qwen3.5-122B-A10B19.6%estimated ± 4.1 pp, medium confidence
81Qwen3.5 Plus19.6%estimated ± 4.1 pp, medium confidence
82Qwen3.7 Plus19.6%estimated ± 4.1 pp, medium confidence
83GLM-519.2%estimated ± 5.6 pp, medium confidence
84Gemini 3 Flash18.8%measured
85GPT-5.418.6%estimated ± 5.6 pp, medium confidence
86Claude Sonnet 4.618.5%estimated ± 4.1 pp, medium confidence
87GLM-5.217.5%estimated ± 4.1 pp, low confidence
88Qwen3.7 Flash17.5%estimated ± 4.1 pp, low confidence
89Gemini 3 Pro15.9%measured
90GPT-5.2-Codex15.9%measured
91GPT-5.114.5%measured
92GPT-4o mini14.2%estimated ± 4.1 pp, low confidence
93Qwen3.6 Plus14.2%estimated ± 4.1 pp, low confidence
94GLM-4.713.0%measured
95DeepSeek V3.211.6%measured
96GPT-5.1-Codex-Max11.6%measured
97MiniMax M3.1 Flash6.9%estimated ± 6.6 pp, medium confidence
98MiMo-V2-Pro6.8%estimated ± 6.6 pp, medium confidence
99MiniMax M2.76.8%estimated ± 6.6 pp, medium confidence
100Hy3 Preview6.8%estimated ± 6.6 pp, medium confidence
101Trinity-Large-Thinking6.8%estimated ± 6.6 pp, medium confidence
102Mistral Medium 3.5 128B6.8%estimated ± 6.6 pp, medium confidence
103Gemini 3.1 Flash-Lite6.8%estimated ± 6.6 pp, medium confidence
104Llama 4 Maverick6.8%estimated ± 6.6 pp, low confidence
105Mistral Large 36.8%estimated ± 6.6 pp, low confidence
106Claude Haiku 4.55.8%measured
107Grok 44.3%measured
108Qwen3 235B 25073.5%estimated ± 7.7 pp, low confidence
109Claude 3.5 Sonnet3.1%estimated ± 7.7 pp, low confidence
110Claude 3.5 Haiku3.1%estimated ± 7.7 pp, low confidence
111Gemini 2.0 Flash3.1%estimated ± 7.7 pp, low confidence
112Grok 4.1 Fast2.9%measured

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General