ToolBench - WebShop Short: leaderboard

ToolBench subtask measuring shorter WebShop shopping-agent performance.

Metric: Task Score. Source: huggingface.co. Saturation forecast: Estimated already saturated. 44 models tracked.

Top models

#ModelScoreOverall rank
1LLaMA-65B41.2#1266
2Llama 2 13B31.67#1414
3LLaMA-30B30.6#1397
4Llama 2 70B30.45#1044
5LLaMA-13B27.6#1495
6CodeLlama-34B-Instruct-hf25.99#1355
7CodeLlama-13B-Instruct-hf9.7#1513
8Llama 2 7B6.92#1499
9CodeLlama-34B-Python-hf6.47#1539
10CodeLlama-34B-hf5.53#1501
11pythia-1.4B5.2#1605
12CodeLlama-13B-Python-hf2.4#1606
13GPT-3.5 Turbo1.8#857
14CodeLlama-7B-Python-hf1.58#1588
15GPT-NeoXT-Chat-Base-20B0.7#1566

No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=toolbench-webshop-short · How It Works · Data refreshed daily, snapshot 2026-11-06.