MiniCPM5-1B: Modelo de 1B de parâmetros executa agentes locais no telemóvel

A OpenBMB lançou o MiniCPM5-1B, um modelo de IA com mil milhões de parâmetros, concebido para execução local em hardware com recursos limitados, agora disponível no Hugging Face. O modelo obtém uma média de 42,57 em benchmarks de agentes e de raciocínio, superando o concorrente seguinte de classe 1B, com 35,61. O MiniCPM5-1B suporta o Model Context Protocol (MCP) e chamadas nativas a ferramentas, permitindo fluxos de trabalho de agentes locais em dispositivos do consumidor sem conectividade à cloud. O modelo cabe nas limitações de memória de um smartphone, mantendo uma janela de contexto de 128K tokens — aproximadamente 96.000 palavras de texto contínuo numa única passagem.

Technical Architecture

MiniCPM5-1B builds on the architectural backbone of MiniCPM4, developed by teams at THUNLP, Tsinghua University, and ModelBest. The core innovation is InfLLM v2, a trainable attention mechanism that processes each token against fewer than 5% of surrounding tokens during long-context inference, reducing computation without meaningful accuracy loss.

The training pipeline introduced UltraClean, a filtering system that achieved competitive performance using 8 trillion training tokens—compared to 36 trillion consumed by Qwen 3. Post-training applied reinforcement learning combined with efficient distillation techniques, raising benchmark scores on math, code, and instruction-following by 16 points while reducing runaway-length responses by 29 percentage points.

Agentic Capabilities and Use Cases

Testing confirmed MiniCPM5-1B supports both MCP and tool calling, placing it on a short list of sub-2-billion-parameter models capable of local agentic workflows without cloud infrastructure. Practical deployment scenarios include local agents on mobile devices that query calendars, search local databases, or call web research MCP servers entirely offline.

The 128K-token context window enables persistent memory across extended interactions—sufficient for roleplay sessions spanning dozens or hundreds of exchanges, document digestion, or multi-step agent tasks without context reset.

Benchmark Performance

OpenBMB's capability benchmark compares MiniCPM5-1B against Alibaba's Qwen3-0.6B, Qwen3.5-0.8B, and Liquid AI's LFM2.5-1.2B-Thinking across seven categories: general knowledge, domain knowledge, coding, instruction-following, math reasoning, logical reasoning, and agentic tasks. MiniCPM5-1B leads across all seven, with the most pronounced margins in agentic performance and general knowledge.

Testing Results

Three evaluations were conducted:

Logic Trap Test: When asked whether it is legal for a man to marry his widow's sister according to Falkland Islands law, the model produced a detailed breakdown of marital law and missed the logical trap—that a man with a widow is deceased. The model treated it as a straightforward jurisdictional question rather than recognizing the logical impossibility.

A/B Choice Test: When asked to determine which industry—Crypto or AI—would dominate the economy in 2100, the model hedged into a both-sides answer rather than reasoning decisively. This represents a known failure mode across small models under conversational pressure.

Tool Calling Test: When asked for the current Bitcoin price and three stock recommendations, the model successfully called the tool. Recommendations provided were Amazon, Microsoft, and Nvidia.

Pairing MiniCPM5-1B with an MCP server for web research substantially mitigates hallucination on obscure factual questions.

Disponibilidade

O MiniCPM5-1B está disponível no Hugging Face sob uma licença Apache 2.0. O modelo é compatível com vLLM, SGLang e frameworks de inferência standard do Transformers. Os utilizadores que necessitem de funcionalidades agentic têm de configurar definições adicionais disponíveis no repositório Github do modelo.