general benchmark
HumanEvalFIM-Average
Average evaluation of HumanEval Fill-in-the-Middle benchmark variants (single-line, multi-line, random-span) for assessing code infilling capabilities of language models
Updated Aug 7, 2026
Published models1
Registry coverage1
MetricScore
EvidenceB
HumanEvalFIM-Average leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MA | 91.6% | 100.0% | 1 | C |
HumanEvalFIM-Average highlights
The top published results on this benchmark's own scale.
What is HumanEvalFIM-Average?
Definition and scoring fields from the benchmark registry.
Average evaluation of HumanEval Fill-in-the-Middle benchmark variants (single-line, multi-line, random-span) for assessing code infilling capabilities of language models
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- HumanEvalFIM-Average
- Modality
- text
- Primary category
- general
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- humanevalfim-average|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about HumanEvalFIM-Average.
Which model scores highest on HumanEvalFIM-Average?
Codestral-22B is currently ranked first with 91.6%.
What does HumanEvalFIM-Average measure?
Average evaluation of HumanEval Fill-in-the-Middle benchmark variants (single-line, multi-line, random-span) for assessing code infilling capabilities of language models
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 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.