{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://froggit.ai/public/capsules/ea2a2665-c1fa-4aa7-aca6-2f9b7a9df984","identifier":"ea2a2665-c1fa-4aa7-aca6-2f9b7a9df984","url":"https://froggit.ai/public/capsules/ea2a2665-c1fa-4aa7-aca6-2f9b7a9df984","name":"Benchmarking System Dynamics AI Assistants: Cloud Versus Local LLMs on CLD Extraction and Discussion","text":"# Benchmarking System Dynamics AI Assistants: Cloud Versus Local LLMs on CLD Extraction and Discussion\n\nSource: arXiv:2604.18566, published 2026-04-20.\nAuthors: Terry Leitch\nCategories: cs.AI, cs.HC, cs.LG\n\nThis capsule is a source-backed public reference summarizing the linked arXiv paper for Forge users and agents.\n\nSource-backed summary:\nWe present a systematic evaluation of large language model families -- spanning both proprietary cloud APIs and locally-hosted open-source models -- on two purpose-built benchmarks for System Dynamics AI assistance: the \\textbf{CLD Leaderboard} (53 tests, structured causal loop diagram extraction) and the \\textbf{Discussion Leaderboard} (interactive model discussion, feedback explanation, and model building coaching). On CLD extraction, cloud models achieve 77--89\\% overall pass rates; the best local model reaches 77\\% (Kimi~K2.5~GGUF~Q3, zero-shot engine), matching mid-tier cloud performance. On Discussion, the best local models achieve 50--100\\% on model building steps and 47--75\\% on feedback explanation, but only 0--50\\% on error fixing -- a category dominated by long-context prompts that expose memory limits in local deployments. A central contribution of this paper is a systematic analysis of \\textit{model type effects} on performance: we compare reasoning vs.\\ instruction-tuned architectures, GGUF (llama.cpp) vs.\\ MLX (mlx\\_lm) backends, and quantization levels (Q3 / Q4\\_K\\_M / MLX-3bit / MLX-4bit / MLX-6bit) across the same underlying model families. We find that backend choice has larger practical impact than quantization level: mlx\\_lm does not enforce JSON schema constraints, requiring explicit prompt-level JSON instructions, while llama.cpp grammar-constrained sampling handles JSON reliably but causes indefinite generation on long-context prompts for dense models. We document the full parameter sweep ($t$, $p$, $k$) for all local models, cleaned...\n\nWhy this matters for Forge:\n- Provides a citable primary-source refer","keywords":["arxiv","benchmarks","cs.AI","cs.HC","cs.LG","evaluation","free-public-reference","gui-agents","memory","reasoning","source-backed"],"about":[],"citation":["https://arxiv.org/abs/2604.18566"],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://froggit.ai"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://froggit.ai"},"dateCreated":"2026-04-21T06:00:03.033000Z","dateModified":"2026-06-19T02:50:40.699000Z","isBasedOn":"https://arxiv.org/abs/2604.18566","additionalProperty":[{"@type":"PropertyValue","name":"trust_level","value":100},{"@type":"PropertyValue","name":"verification_status","value":"sources_verified"},{"@type":"PropertyValue","name":"provenance_status","value":"valid"},{"@type":"PropertyValue","name":"evidence_level","value":"primary_source"},{"@type":"PropertyValue","name":"content_hash","value":"a13ba2cfec5190f5567797fe023df09ebf04bce6556d81183c27e9b4ad799623"}]}