Tool calling (internal eval)

Tool calling (internal eval) is a benchmark for handling several input types together (images, documents, text). The official description says: "Tool calling (internal eval) as evaluated on the LFM3-Omni-2B card: Liquid AI internal voice tool-calling evaluation, audio input, percentage, higher is better". Below we describe in simple words what benchmarks of this type generally examine.

Tool calling (internal eval) measures handling several input types together (images, documents, text). In detail, it checks:

• questions that mix text with images or document pages
• reading and extracting information from screenshots and scans
• understanding layout, titles, logos and figures in real documents
• staying accurate when the text content alone is not enough

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better.

Official methodology note: Tool calling (internal eval) as evaluated on the LFM3-Omni-2B card: Liquid AI internal voice tool-calling evaluation, audio input, percentage, higher is better.

What is the Tool calling (internal eval) benchmark?

Tool calling (internal eval) is a benchmark for handling several input types together (images, documents, text). The official description says: 'Tool calling (internal eval) as evaluated on the LFM3-Omni-2B card: Liquid AI internal voice tool-calling evaluation, audio input, percentage, higher is better'. Below we describe in simple words what benchmarks of this type generally examine.

What does Tool calling (internal eval) test?

Tool calling (internal eval) mainly measures handling several input types together (images, documents, text). Concretely, it checks: questions that mix text with images or document pages; reading and extracting information from screenshots and scans; understanding layout, titles, logos and figures in real documents.

How is the Tool calling (internal eval) score calculated?

The score is accuracy: the percentage of tasks answered or solved correctly. 100% means every task was solved perfectly; higher is better. The declared maximum score of this benchmark is 100.0.

Where do I get Tool calling (internal eval) data and scores?

The official benchmark page/reference is linked on this page. Every score on Open LLM Radar is also available as open data: download the scores JSON (http://open-llm-radar-api:8000/api/v1/benchmarks/tool-calling-internal-eval/scores) or pull it directly from the Open LLM Radar API with any HTTP client.