FrogNano-4B-2609 (GGUF)

GGUF quantizations of microsoft/FrogNano-4B-2609 — Microsoft's 4B repository-level coding agent derived from Qwen3.5-4B and post-trained only with reinforcement learning on synthetic software-engineering tasks — optimized for local inference with llama.cpp and Ollama, with its vision projector. All quants use an importance matrix.

Quick Start

With Ollama

ollama run richardyoung/frognano-4b-2609
# or straight from Hugging Face
ollama run hf.co/richardyoung/FrogNano-4B-2609-GGUF:Q4_K_M

With llama.cpp

huggingface-cli download richardyoung/FrogNano-4B-2609-GGUF \
    --include "*Q4_K_M*" "mmproj*" --local-dir ./models

./llama-mtmd-cli -m ./models/FrogNano-4B-2609-Q4_K_M.gguf \
    --mmproj ./models/mmproj-FrogNano-4B-2609-F16.gguf \
    --image page.png -p "grounding"

Microsoft evaluated it with its five-tool Leaf harness; the vision tower is inherited from Qwen3.5-4B and was not post-trained.

Why This Model?

• Repository-level coding agent in 4B
• Post-trained purely with RL — no stronger-model trajectories
• Thinking model, 262K-token context
• MIT license

Use Cases

• Local coding agents
• Code review and repo Q&A
• Research on RL-trained agents

Available Quantizations

Quant Quality

Each quant compared with the F16 model using llama-perplexity --kl-divergence on the wikitext-2 test set (40 × 512-token chunks). KL divergence measures how far the quant's next-token probabilities drift from F16; lower is better. Top-token match is how often the quant picks the same next token as F16. As a rough guide, a KLD under 0.01 is hard to tell from F16, and from about 0.05 differences start to show in longer outputs.

Verified

Tested through Ollama before upload (Q4_K_M and Q8_0) on a synthetic invoice with the prompt None; Q4_K_M returned: 'INVOICE 4471 / Date: 2026-09-28 / Widgets x3 ....... $42.00 / TOTAL: $86.20'.

Also Available

Other Models by richardyoung

• New Heretic models (reproducible, listed in the Heretic Grimoire): LFM2.5-1.2B | Granite-4.2-3B | Granite-4.2-8B | Olmo-3-7B-Think | Phi-4-mini | DeepSeek-R1-0528-Qwen3-8B | Agents-A1-4B | FrogNano-4B | Muse-Glimmer-30B | Hemmingway-1
• Abliterated/Uncensored models: Qwen2.5-7B | Qwen3-14B | DeepSeek-R1-32B | Qwen3-8B
• OCR & Vision: MiniCPM-V-4.6 GGUF | dots.mocr GGUF | Unlimited-OCR GGUF | olmOCR GGUF | LightOnOCR-3-4B GGUF | WeVisDoc-4B GGUF | Xiaomi-OCR-0 GGUF
• MLX quantizations (Apple Silicon): Kimi-K2 series | olmOCR MLX
• Healthcare/Medical: Synthea 575K patients dataset | CardioEmbed
• Research: LLM Instruction-Following Evaluation (arxiv:2510.18892)

---

Built & maintained by Richard Young · DeepNeuro