reasoning benchmark
Gorilla Benchmark API Bench
APIBench, a comprehensive dataset of over 11,000 instruction-API pairs from HuggingFace, TorchHub, and TensorHub APIs for evaluating language models' ability to generate accurate API calls.
Updated Aug 7, 2026
Published models3
Registry coverage3
MetricScore
EvidenceB
Gorilla Benchmark API Bench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | ME | 35.3% | 100.0% | 3 | C | |
| 2 | ME | 29.7% | 50.0% | 3 | C | |
| 3 | ME | 8.2% | 0.0% | 3 | C |
Score distribution
Top published rows on the benchmark's original scale.
Gorilla Benchmark API Bench highlights
The top published results on this benchmark's own scale.
What is Gorilla Benchmark API Bench?
Definition and scoring fields from the benchmark registry.
APIBench, a comprehensive dataset of over 11,000 instruction-API pairs from HuggingFace, TorchHub, and TensorHub APIs for evaluating language models' ability to generate accurate API calls.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- Gorilla Benchmark API Bench
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- gorilla-benchmark-api-bench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about Gorilla Benchmark API Bench.
Which model scores highest on Gorilla Benchmark API Bench?
Llama 3.1 405B Instruct is currently ranked first with 35.3%.
What does Gorilla Benchmark API Bench measure?
APIBench, a comprehensive dataset of over 11,000 instruction-API pairs from HuggingFace, TorchHub, and TensorHub APIs for evaluating language models' ability to generate accurate API calls.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.