StarCoderBase-3B: benchmark results
Provider: BigCode. Access: Open.
Aggregate Rating 1100 ± 35, rank #10726 of 10916 rated models, from 10 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 43.05 | MC2 (%) (0-shot) | 22.4 |
| Open LLM Leaderboard v1 - GSM8K | 1.74 | Accuracy (%) (5-shot) | 18.2 |
| Open LLM Leaderboard v1 - MMLU | 27.35 | Accuracy (%) (5-shot) | 15.1 |
| EvalPlus (HumanEval+ & MBPP+) | 15.9 | Pass@1 avg (%) | 10.9 |
| BigCode Models Leaderboard | 21.5 | HumanEval Python Pass@1 (%) | 10.2 |
| Open LLM Leaderboard v1 - HellaSwag | 39.11 | Normalized accuracy (%) (10-shot) | 7.9 |
| Open LLM Leaderboard v1 - ARC Challenge | 25.85 | Normalized accuracy (%) (25-shot) | 5.2 |
| Open LLM Leaderboard v1 - WinoGrande | 51.14 | Accuracy (%) (5-shot) | 5 |
| CanItEdit (Descriptive) | 14.81 | pass@1 (%; share of the 105 hand-written Python editing prob | 4.2 |
| CanItEdit (Lazy) | 9.9 | pass@1 (%; share of the 105 hand-written Python editing prob | 4.2 |
Interactive version: theaggregate.ai/model?slug=starcoderbase-3b · How It Works · Data refreshed daily, snapshot 2026-11-08.