Tag

SVG Generation Benchmark (Static)

Compares 30 frontier LLMs generating static SVG markup from 500 prompts using 1,355,161 human votes across three leaderboards (Preference, Coherence, Alignment); provides raw SVGs, 768×768 rasterized PNGs, and per-comparison human vote records under a CC-BY-4.0 prompt license.

Ornith-1.0-397B

Provides an open-source Mixture-of-Experts coding LLM (397B) optimized for agentic, tool-enabled coding workflows with a 262,144-token context window, OpenAI-compatible API, serving recipes (vLLM/SGLang), and published coding-benchmark results.

LFM2.5-230M

230M-parameter multilingual instruction-tuned text-only LLM for on-device agentic pipelines and data extraction; 32K context, 19T-token pretraining, optimized for fast CPU/edge inference (e.g., 213 tok/s on Galaxy S25 Ultra, 42 tok/s on Raspberry Pi 5); not for heavy reasoning or complex code generation.

EdgeBench

Measures how autonomous AI agents learn via long-horizon, feedback-rich executable tasks; publishes 51 public tasks from a 134-task suite and provides SForge, a two-container evaluation harness for iterative 12+ hour runs to track learning trajectories.

IFStruct v1.0

Measures whether models produce valid JSON/YAML that strictly follow a requested schema across diverse, naturally phrased prompts. Contains 2,000 frozen test prompts with binary structural validation (no constrained decoding), focusing on schema compliance and edge cases like escaping, wrapper keys, and fenced code blocks.

Ornith-1.0-35B-GGUF

A self-improving, agentic coding LLM tailored for terminal-style coding agents and tool-calling, provided as 35B MoE GGUF weights with very large context support. Trained with reinforcement learning to jointly generate task scaffolds and solutions; designed for local inference and OpenAI-compatible tool endpoints.

Ornith-1.0-9B-GGUF

Provides a GGUF-quantized local build of Ornith-1.0's 9B dense model for offline inference and terminal-focused coding agents. Supports OpenAI-compatible tool-calling, a 256K context window, and runs via llama.cpp or Ollama on a single high-memory GPU.

DeepSeek-V4-Pro-DSpark

Mixture-of-Experts LLM designed for million-token contexts, combining hybrid compressed attention, FP4/FP8 quantization-aware training for MoE experts, and multi-mode 'thinking' (Non-think/Think High/Think Max); includes a speculative-decoding extension for faster inference.

DeepSeek-V4-Flash-DSpark

A Hugging Face model checkpoint that attaches a speculative decoding module to DeepSeek-V4-Flash, enabling million-token context handling with MoE architecture, FP4/FP8 mixed precision, and long-context inference optimizations.

The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning

Proposes Monotonic Inference Policy Improvement (MIPI) and a two-step Monotonic Inference Policy Update (MIPU) to address training–inference probability mismatch in LLM reinforcement learning by constructing sampler-referenced candidate updates and accepting synchronized updates using an inference-gap proxy; shows improved reasoning accuracy and stability under FP8-quantized rollouts.

Huihui-GLM-5.2-abliterated-GGUF

An uncensored GGUF build of GLM-5.2 that applies weight “abliteration” to remove refusal filters and produce a locally runnable text-generation model; includes quantization conversions and shard-merge instructions, intended for experimental research rather than production use.

Metacognition-Bench

Provides 300 adversarial "metacognitive-trap" problems to measure whether LLMs notice and recover from their own reasoning errors. Combines multiple-choice vulnerability tests with free-form adapter-gain evaluation and ships per-model metacognition adapters for frozen-base probing.