Skip to content

Benchmark

Open Agent Leaderboard AppWorld test_normal

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
5
Scores recorded
25
Sources
1
IQ weight
0.92

Score distribution

26%
Mean
11%
Median
25.5%
Std dev
0%
10th percentile
66%
90th percentile
0%
Min
0%70%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.92 for chat models, measured over 5 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 4.5agent: smolagents_code, task: appworld_test_normal, subset: test_normal, route: openai/aws· Open Agent Leaderboard results data docs70%
2Claude Opus 4.5agent: openai_solo, task: appworld_test_normal, subset: test_normal, route: openai/aws· Open Agent Leaderboard results data docs68%
3Claude Opus 4.5agent: claude_code, task: appworld_test_normal, subset: test_normal, route: openai/aws· Open Agent Leaderboard results data docs66%
4Claude Opus 4.5agent: tool_calling_with_shortlisting, task: appworld_test_normal, subset: test_normal, route: openai/aws· Open Agent Leaderboard results data docs64%
5Claude Opus 4.5agent: tool_calling, task: appworld_test_normal, subset: test_normal, route: openai/aws· Open Agent Leaderboard results data docs61%
6Gemini 3 Pro Previewagent: openai_solo, task: appworld_test_normal, subset: test_normal, route: openai/gcp· Open Agent Leaderboard results data docs57%
7Gemini 3 Pro Previewagent: tool_calling_with_shortlisting, task: appworld_test_normal, subset: test_normal, route: openai/gcp· Open Agent Leaderboard results data docs55%
8Gemini 3 Pro Previewagent: tool_calling, task: appworld_test_normal, subset: test_normal, route: openai/gcp· Open Agent Leaderboard results data docs50%
9Gemini 3 Pro Previewagent: claude_code, task: appworld_test_normal, subset: test_normal, route: openai/gcp· Open Agent Leaderboard results data docs36%
10GPT 5.2agent: tool_calling_with_shortlisting, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs22%
11DeepSeek V3.2agent: smolagents_code, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs13%
12Gemini 3 Pro Previewagent: smolagents_code, task: appworld_test_normal, subset: test_normal, route: openai/gcp· Open Agent Leaderboard results data docs13%
13Kimi K2.5agent: smolagents_code, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs11%
14Kimi K2.5agent: tool_calling_with_shortlisting, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs10%
15DeepSeek V3.2agent: tool_calling, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs9%
16Kimi K2.5agent: tool_calling, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs9%
17Kimi K2.5agent: claude_code, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs8%
18Kimi K2.5agent: openai_solo, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs8%
19GPT 5.2agent: smolagents_code, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs7%
20DeepSeek V3.2agent: openai_solo, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs6%
21DeepSeek V3.2agent: tool_calling_with_shortlisting, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs4%
22DeepSeek V3.2agent: claude_code, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs3%
23GPT 5.2agent: claude_code, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs0%
24GPT 5.2agent: openai_solo, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs0%
25GPT 5.2agent: tool_calling, task: appworld_test_normal, subset: test_normal, route: openai/azure· Open Agent Leaderboard results data docs0%

Data credits: Open Agent Leaderboard results (Exgentic), CDLA-Permissive-2.0; agreement text at https://cdla.dev/permissive-2-0/. Fractions shown as percentages. (licence)