Skip to content

Benchmark

AppWorld Test-Normal: Scenario Goal Completion

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

Models and systems measured
26
Scores recorded
26
Sources
1
IQ weight
No weight published

Score distribution

41.4%
Mean
32.1%
Median
31.2%
Std dev
8.9%
10th percentile
87.5%
90th percentile
0%
Min
0%91.1%
The band covers the middle 80% of scores. The line marks the median.

Results

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Claude Opus 4.6Button Agent (f072cd07-6ae0fbdf)· AppWorld Leaderboard data docs89.3%
claude-opus-4-7LARA MAS (23715209-d5c9ade3)· AppWorld Leaderboard data docs82.1%
DeepSeekCoderFullCodeRefl (0465049f-bff9fca2)· AppWorld Leaderboard data docs8.9%
DeepSeekCoderPlanExec (871bf9fd-4de66f57)· AppWorld Leaderboard data docs0%
DeepSeekCoderReAct (9ae299b8-25c255f5)· AppWorld Leaderboard data docs1.8%
GLM-5.3AtlasHarness (65df675b-ca5e4771)· AppWorld Leaderboard data docs87.5%
GLM-5.3-FlashIBM CUGA (a2db3f1e-2c28b1b2)· AppWorld Leaderboard data docs91.1%
GPT-4 TurboFullCodeRefl (571aa685-a8671c57)· AppWorld Leaderboard data docs19.6%
GPT-4 TurboIPFunCall (08d2ff3e-a0589430)· AppWorld Leaderboard data docs21.4%
GPT-4 TurboPlanExec (c722a232-74b9384b)· AppWorld Leaderboard data docs16.1%
GPT-4 TurboReAct (1f0055c0-123849bb)· AppWorld Leaderboard data docs12.5%
GPT-4.1IBM CUGA (8a02d49a-0df800d6)· AppWorld Leaderboard data docs62.5%
gpt-4.1-miniLARA MAS (695a4b55-1b68dde4)· AppWorld Leaderboard data docs50%
GPT-4oFullCodeRefl (d035381e-2d30c0e3)· AppWorld Leaderboard data docs26.8%
GPT-4oIPFunCall (3a889687-923696fe)· AppWorld Leaderboard data docs16.1%
GPT-4oPlanExec (ac5b742b-aa0243da)· AppWorld Leaderboard data docs23.2%
GPT-4oReAct (3414a75f-3df6a61f)· AppWorld Leaderboard data docs32.1%
GPT-4oReAct + 2 SetBSR Demos (1c12639a-57a0ac34)· AppWorld Leaderboard data docs57.1%
gpt-5.6-lunakecaipan capybara (1095b768-6513e0c6)· AppWorld Leaderboard data docs73.2%
gptoss-20bRooAGI_agentbuilder_memtune_rl (5acab4bb-badb978f)· AppWorld Leaderboard data docs58.9%
LLaMA3-70BFullCodeRefl (baf95530-cee120fc)· AppWorld Leaderboard data docs17.9%
LLaMA3-70BPlanExec (77f25849-862250c4)· AppWorld Leaderboard data docs1.8%
LLaMA3-70BReAct (4077db41-07bc5204)· AppWorld Leaderboard data docs8.9%
Qwen2.5-32BLOOP (e5e08b7e-d8e6474f)· AppWorld Leaderboard data docs53.6%
Qwen3-14BAlibaba Cloud ApsaraLab AgentRL (cabab065-0caee3c3)· AppWorld Leaderboard data docs80.4%
Qwen3.5-27BSynthAPI (3bd98253-d4fa6c8e)· AppWorld Leaderboard data docs83.9%

Data credits: AppWorld leaderboard (Stony Brook NLP). Aggregate scores from the official leaderboard; no results licence asserted. Protected bundles and experiment outputs are not used