MiniCPM5-1B: benchmark results
A dense 24-layer LlamaForCausalLM checkpoint with 1,080,632,832 total parameters, including 679,552,512 non-embedding parameters, and native 131,072-token context. The publisher explicitly supports Think and No Think modes through the same post-trained checkpoint; base and SFT-only releases remain separate. Provider: OpenBMB. Released 2026-05-21. Access: Open.
Aggregate Rating 1267 ± 28, rank #1497 of 1635 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Tau2-Bench Telecom | 82.46 | Success Rate (%) | 73.3 |
| AA TAU-2 Bench | 82.46 | Accuracy (%) | 72.9 |
| Kurakurai French LLM - AIME 2024 | 2.67 | Accuracy (%) | 67.1 |
| Kurakurai French LLM - AIME 2025 | 2.67 | Accuracy (%) | 67.1 |
| Kurakurai French LLM - HumanEval+ | 42.44 | pass@1 (%) | 65.9 |
| AA IFBench | 49.32 | Accuracy (%) | 58.7 |
| Kurakurai French LLM - MultiIF | 37.16 | Accuracy (%) | 53.7 |
| Kurakurai French LLM - MATH-500 | 33.96 | Accuracy (%) | 46.3 |
| Kurakurai French LLM - MMLU-Pro | 19.5 | Accuracy (%) | 41.5 |
| Kurakurai French LLM - MBPP | 30.45 | pass@1 (%) | 39 |
| AA Humanity's Last Exam | 6.49 | Accuracy (%) | 38.1 |
| Kurakurai French LLM - IFEval | 48.21 | Accuracy (%) | 34.1 |
Interactive version: theaggregate.ai/model?slug=minicpm5-1b · How It Works · Data refreshed daily, snapshot 2026-11-05.