NeoHorse

NanoJev is an open-source 0.6B parallel-decision model and training pipeline based on Qwen3-0.6B. It accepts states, questions, and candidate sets in a single forward pass and returns complete probability distributions without output-token decoding. Shared decision heads support dynamic-choice distributions over supplied candidates, Boolean proposition probabilities, and ordered score levels with expected values. The repository includes data generation, distribution-loss training, evaluation, persistent serving, and browser replay tools. Its game demonstrations combine model judgments with controller code: maze controllers ask local safety questions and track collisions and explored paths, while Snake controllers filter immediate collisions, plan toward visible food, and use the model to break ties between remaining actions. Checkpoints and datasets are distributed through Hugging Face, and the service exposes repeated batched evaluations through a local HTTP endpoint.

LightNav-0 is an open-source generalist embodied-navigation model from the Light Origins Team. It uses a pretrained Qwen3-VL backbone with a single egocentric RGB stream and natural-language instructions to control humanoid, quadruped, wheeled, and aerial robots, transferring across tasks, embodiments, and scenes without per-task or per-benchmark fine-tuning. The model extends the vocabulary with dual-channel pointing tokens and residual vector-quantized action tokens rather than adding navigation-specific prediction heads. At each step, it emits an affordance point representing a feasible local direction or waypoint and an object point representing the goal, followed by three action tokens that decode to ten future SE(2) waypoints for an embodiment-specific low-level controller. Its temporally aware history compressor samples older frames less frequently, pools them more coarsely, and preserves ordering with timestamp tokens while bounding the visual context. The repository includes a released checkpoint and action decoder, command-line prediction, a vLLM-based server with WebSocket streaming, a MuJoCo simulation demo, evaluation harnesses, and a ROS 2 deployment stack with adapters for the Unitree Go2 and LimX TRON 1. The project is released under the Apache License 2.0; the included EVT-Bench is separately licensed CC BY-NC-SA 4.0.

Bespoke Nimble is an open-source model, training recipe, and local scoring library from Bespoke Labs for making typed decisions about text. Given a context and a flat schema of boolean or fixed-choice fields, it selects an allowed answer and returns candidate logits and probabilities without generating a reasoning explanation or free-form answer. Nimble fine-tunes Qwen3.5-9B with LoRA using contrastive training pairs: two nearly identical examples differ in one relevant fact, which changes the correct label. At inference, the scorer reads the logits for one-token candidate codes and converts them to probabilities with softmax; the Python library constructs the typed result rather than parsing generated JSON. The MLX ParallelScorer processes shared context once and scores fields in parallel, while the CUDA scorer scores each field separately. The repository includes the model recipe, curation and evaluation code, datasets, examples, tests, and deployment documentation. Bespoke-Nimble-9B can run locally on Apple Silicon or an NVIDIA BF16-capable GPU. It accepts text only, does not support nested schemas or answers outside the supplied choices, and its probabilities are not guarantees of correctness.

oh-my-hermes (OMH) is a plugin and operating layer for Hermes Agent that adds model routing, coding workflows, specialist skills, project memory, and evidence-gated execution without replacing Hermes as the natural-language interface. It scores requests, selects workflows and model-and-effort categories, applies model-family-specific prompting, and can split accepted plans into parallel worktree-based lanes with typed results and verification gates. Its workflows cover interviewing, research, planning, coding delegation, quality assurance, performance work, and iterative plan-build-review loops; the terminal interface exposes phase-based todos, delegated-lane status, cost and token telemetry, and execution states that distinguish preparation, reported completion, and verified results. OMH also provides a file-backed, reviewer-gated memory store with provenance, review dates, conflict handling, and task-scoped recall packs, while leaving Hermes's native memory untouched. It is distributed as a command-line package and plugin with installation paths including a shell installer, Homebrew, Bun, npm, and Hermes skill tap.

HarnessRouter Community Edition is the open-source edition of HarnessRouter, an agent-harness infrastructure that provides a unified Agent API to integrate language models and agent runtimes. It is maintained by HarnessRouter and is intended for deployment on users' own infrastructure to connect models such as Codex, Claude Code, and Hermes.