BFCL V4

Reported in recent model launches

These model reports include this benchmark. Their harnesses, task subsets, scoring policies, and resource budgets can differ; follow each report’s evaluation notes before joining the numbers.

What it measures

Berkeley’s function-calling evaluation covers tool invocation and broader agentic interactions.

How to read the result

Inspect subcategories to separate argument construction, relevance decisions, multi-turn behavior, and longer interactions.

Limits and comparison traps

Native function calling and prompt-based tool calling expose different interfaces. Report benchmark version and model adapter.

Benchmark reference notes

Comprehensive function-calling evaluation spanning AST matching, live API execution, multi-turn, multi-step, relevance, and agentic web/memory calls.

Epoch data and evaluation setting

Epoch AI publishes this benchmark in its capability collection. Imported rows retain the published score column and model configuration. Original authors and evaluation providers remain credited in the result references.

Epoch attribution and score coverage

Data is attributed to Epoch AI under CC BY 4.0, with original evaluation providers credited in the linked source snapshots. Scores are imported only where a published mapping identifies the field and its scale; missing or incompatible records are omitted rather than assigned zero.

Where this fits in an evaluation suite

A transcript that says “done” is weaker evidence than the verified final state. Track policy violations, recoveries, and invalid attempts alongside success.

Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.

A reproducible comparison

Pin the benchmark release, task split, model configuration, tools, sampling budget, and evaluation harness described by the primary source.

Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.