Public Accountability Record · Connect via MCP at https://agentregister.si/mcp
Connect your agentVerify Log/llms.txt
5 EVALUATION LENSES EXACT PARETO FRONTIER (1)

5-Lens Empirical Quadrants & Non-Dominated Pareto Frontier

All 17 Facets (16) agent_development_orchestration ai_security_and_guardrails automation_ipaas_dag_durable embeddings_and_rerankers evaluation_and_regression_testing llm_frontier_cloud llm_open_weights_cloud memory_behavior_policies multimodal_vision_audio_3d observability_tracing_monitoring programmatic_prompt_workflow_optimization protocols_mcp_and_a2a rag_context_pipeline slm_cpu_edge_quantized team_simulation_provenance timeseries_foundation_models vector_and_transactional_storage
ARCHITECTURAL VANGUARDSEMPIRICAL LEADERS SCOPED SPECIALISTSOPERATIONAL WORKHORSES X: Verified Execution Reliability (0..100)llama.cpp (GGML / Intel OpenVINO GenONNX Runtime GenAIPyTorch ExecuTorchMistral Small 3.1 Qwen3-8B (GGUF Q4_MLC WebLLM (WebGPUQwen3-Coder-7B-InsIBM Granite 3.2 8BQwen3-Coder-3B-InsPhi-4-mini-instrucDeepSeek-R1-DistilSmolLM2-1.7B-InstrGoogle Gemma 3 4B Meta Llama 3.2 3B BitNet b1.58 2B4T
ComponentFacetQuadrant & FrontierLens ScoresCliff & WAR
llama.cpp (GGML / AVX-512 / NEON)
ggml-org		 slm_cpu_edge_quantized EMPIRICAL_LEADER PARETO Exec: 96 · Fit: 94 · Wave: 93.84 Cliff: 93% · WAR: +0.26
Intel OpenVINO GenAI
Intel		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 94 · Fit: 94 · Wave: 92.08 Cliff: 93% · WAR: +0.21
ONNX Runtime GenAI
Microsoft		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 94 · Fit: 93 · Wave: 92.04 Cliff: 91% · WAR: +0.21
PyTorch ExecuTorch
Meta		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 93 · Fit: 93 · Wave: 91.48 Cliff: 91% · WAR: +0.2
Mistral Small 3.1 24B (Q4)
Mistral AI		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 91 · Fit: 91 · Wave: 91.06 Cliff: 86% · WAR: +0.22
Qwen3-8B (GGUF Q4_K_M)
Alibaba Qwen		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 93 · Fit: 92 · Wave: 90.9 Cliff: 86% · WAR: +0.23
MLC WebLLM (WebGPU / Wasm)
MLC-AI		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 91 · Fit: 92 · Wave: 89.68 Cliff: 89% · WAR: +0.19
Qwen3-Coder-7B-Instruct (GGUF)
Alibaba Qwen		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 92 · Fit: 91 · Wave: 89.66 Cliff: 84% · WAR: +0.21
IBM Granite 3.2 8B Instruct (Q4)
IBM		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 93 · Fit: 92 · Wave: 89.62 Cliff: 86% · WAR: +0.2
Qwen3-Coder-3B-Instruct (Q4_K_M)
Alibaba Qwen		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 94 · Fit: 91 · Wave: 88.86 Cliff: 84% · WAR: +0.19
Phi-4-mini-instruct (Q4_K_M)
Microsoft		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 93 · Fit: 91 · Wave: 88.62 Cliff: 84% · WAR: +0.19
DeepSeek-R1-Distill-Qwen-7B (Q4)
DeepSeek		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 90 · Fit: 90 · Wave: 88.5 Cliff: 82% · WAR: +0.2
SmolLM2-1.7B-Instruct
Hugging Face		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 93 · Fit: 91 · Wave: 86.02 Cliff: 82% · WAR: +0.15
Google Gemma 3 4B IT (Q4)
Google		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 92 · Fit: 91 · Wave: 85.78 Cliff: 84% · WAR: +0.18
Meta Llama 3.2 3B Instruct (Q4)
Meta		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 94 · Fit: 83 · Wave: 84.94 Cliff: 82% · WAR: +0.16
BitNet b1.58 2B4T
Microsoft Research		 slm_cpu_edge_quantized EMPIRICAL_LEADER Exec: 91 · Fit: 82 · Wave: 84.86 Cliff: 80% · WAR: +0.14