No official torrent yet — requesting it tells the flock what to host next. You cannot seed a catalog-only page.
Cataloged from Jackrong/Qwopus-GLM-18B-Merged-GGUF
All models by Jackrong →A 64-layer frankenmerge of two of Jackrong's incredible Qwen3.5-9B finetunes, stacking all 32 layers from each to create an ~18B parameter model, then healed with a 1000-step LoRA fine-tune to smooth the layer boundary.
This was a fun experiment! A lot of people have been asking for something between Jackrong's 27B and 9B models — something that runs well on 12–16 GB GPUs. This frankenmerge is an attempt at filling that gap, and the results are surprisingly good.
[!NOTE] Thanks to the creator of this model, @KyleHessling1 🙌 This is still an experimental model, so it may have quirks or issues. If you run into anything weird, or if you make something cool with it, reach out on X.
The raw frankenmerge had a known issue: garbled code output. Because two separately-trained models were stacked at layer 32, structured output (code blocks, HTML, bracket matching) would occasionally come out malformed or hallucinated.
We ran a 1000-step QLoRA heal fine-tune using Jackrong's own training data to let gradients flow across the layer boundary — and the results are significant:
</html> closure.longest_substring sliding-window algorithm test (8/8 cases passing)The healed GGUF (Qwopus-GLM-18B-Healed-Q4_K_M.gguf) is the only version in this repo. If you're interested in the raw unhealed merge for research purposes, reach out.
We ran a 44-test capability suite covering basic generation, reasoning, tool calling, agentic workflows, structured output, context handling, multilingual, programming, and performance.
The healed merge outperforms the brand new Qwen 3.6-35B-A3B MoE (Q4_K_M, 22 GB) despite being significantly smaller (Q4_K_M, 9.2 GB):
[!WARNING] This model has not undergone a complete or comprehensive evaluation. These benchmark results are provided for reference only.
The raw frankenmerge had code formatting issues (garbled code blocks, missing brackets). We ran a 1000-step QLoRA heal fine-tune using Jackrong's training data to smooth the layer-32 boundary:
Jackrong/Qwen3.5-reasoning-700x (70%), Jackrong/Competitive-Programming-python-blend (15%), Jackrong/MultiReason-ChatAlpaca (15%)Three programming tests still fail on the healed version: one function naming issue, one missing JS paren, and one that doesn't produce a code block for pytest generation. These are residual formatting artifacts from the merge.
We put the healed model through a rigorous frontend stress test: 6 increasingly complex HTML/CSS/JS generation tasks, each requiring thousands of tokens of structurally valid code output. The results speak for themselves:
62/63 total checks passed (98.4%)
Every single output had:
The only miss: the Snake game had a minor closing tag typo (html> instead of </html>) at the very end.
This is remarkable for a frankenmerge of two 9B models with only 1000 steps of QLoRA healing. The model is producing production-quality frontend code — not just syntactically valid HTML, but sophisticated interactive applications with modern CSS (Grid, Flexbox, custom properties, keyframe animations) and non-trivial JavaScript (IntersectionObserver, requestAnimationFrame game loops, real-time form validation, SVG chart generation).
All 6 sample HTML files are included in the samples/ directory of this repo — download them and open in a browser to see for yourself.
All credit for the source models goes to Jackrong, who created both of these excellent finetunes. I just stacked the layers — the quality comes from his work.
A reasoning-enhanced finetune of Qwen3.5-9B trained with ~2x more SFT data than v3, focused on structured reasoning, tool-augmented workflows, and multi-step agentic tasks.
Key insight from the v3.5 design: "Scaling high-quality SFT data may further enhance the generalization ability of large language models." Reasoning SFT helps models better utilize existing knowledge and activate latent knowledge through structured reasoning, rather than simply memorizing long Chain-of-Thought outputs.
Performance highlights (27B line reference):
Resources:
A distilled variant of Qwen3.5-9B trained on high-quality reasoning data from a GLM-5.1 teacher model (~700x scale of Qwen3.5-reasoning-700x). Focused on structured reasoning ability, instruction-following consistency, and problem decomposition.
The model learns a structured reasoning scaffold: understand the task, break down the problem, reason step by step, then construct the final answer.
Training data:
Jackrong/GLM-5.1-Reasoning-1M-Cleaned (cleaned from Kassadin88/GLM-5.1-1000000x)Jackrong/Qwen3.5-reasoning-700xResources:
By stacking two differently-distilled reasoning models, this merge combines:
The hypothesis: deeper networks with diverse reasoning training produce more robust, capable models — and the benchmark results suggest it works, at least for the capabilities we tested.
The full BF16 safetensors are not included in this GGUF repo. If you need them for further fine-tuning or experimentation, reach out on X.
Both source models reference: Ren et al., 2026 — Rethinking Generalization in Reasoning SFT (arXiv:2604.06628)
Key findings: reasoning SFT generalizes when sufficiently trained; high-quality long-CoT data enables cross-domain transfer; stronger models learn reasoning structure, not just longer outputs.
This was just for fun — reach out on X if you have questions, find issues, or build something cool with it!
[!WARNING] ⚠️ This is an experimental test model created as part of a fun community exploration project. It is a frankenmerge built without additional post-training or comprehensive safety evaluation. As a result, the model may exhibit unstable behavior, formatting errors, inconsistent reasoning, or other unexpected artifacts. It is intended for research and experimentation only.
Layers 0–31: Jackrong/Qwopus3.5-9B-v3.5 (Opus reasoning distill)Layers 32–63: Jackrong/Qwen3.5-9B-GLM5.1-Distill-v1 (GLM-5.1 reasoning distill)
Embeddings, LM head, MTP, vision encoder: from Qwopus3.5-9B-v3.5llama-server \ -m Qwopus-GLM-18B-Merged-Q4_K_M.gguf \ --chat-template-file your-qwen35-template.jinja \ --ctx-size 65536 \ --flash-attn on \ --n-gpu-layers 99@misc{jackrong_qwopus35_9b_v35, title = {Qwopus3.5-9B-v3.5}, author = {Jackrong}, year = {2026}, publisher = {Hugging Face}}
@misc{jackrong_qwen35_9b_glm51_distill_v1, title = {Qwen3.5-9B-GLM5.1-Distill-v1}, author = {Jackrong}, year = {2026}, publisher = {Hugging Face}}