GGUF iMat

Qwen3.6 27B · Isometry Terse

An isometric merge of Qwen/Qwen3.6-27B with seven reasoning finetunes.

Base Qwen3.6 is a strong thinker with no sense of when to stop — on GSM8K it routinely spends a thousand tokens on a problem it settled in the first hundred, and sometimes runs off the end of a 4k budget without answering at all. This merge answers in a fraction of the reasoning, more accurately, and does not run away.

Model set

Each source was extracted against the base and folded in as its own rotation block, in priority order. Rescale is the magnitude parity factor applied before the fold — source magnitudes span nearly three orders of magnitude, and applied raw the largest of them destroys the model.

Reasoning cost

GSM8K, thinking enabled, 4096-token budget, generated tokens counted directly:

The distributions barely overlap — the merge's worst case runs shorter than a quarter of base's median.

What it is

The body is composed with an isometric merge. Each source delta is reduced to the rotation it induces on the base weight inside the subspace it actually moved, and applied as W → W + Z(R − I)ZᵀW. Because R is orthogonal the merged weight keeps the norm it started with, so there is nothing to renormalize afterward and nothing for a low-bit quantizer to spend dynamic range on.

Sources fold as separate rotation blocks in priority order rather than averaging together, each residualized against the blocks already placed. The AEON overlay is applied after the body, restricted to output projections.

Same operator as the Gemma-4 Isometry models.

Usage

Standard Qwen3.6 chat template, thinking on or off. The efficiency shows with thinking enabled.

from transformers import AutoTokenizer, Qwen3_5ForConditionalGeneration

model_id = "maldv/Qwen3.6-27B-Isometry-Terse"
tok = AutoTokenizer.from_pretrained(model_id)
model = Qwen3_5ForConditionalGeneration.from_pretrained(
    model_id, dtype="bfloat16", device_map="auto")

msgs = [{"role": "user", "content": "..."}]
ids = tok.apply_chat_template(msgs, add_generation_prompt=True,
                              return_tensors="pt", enable_thinking=True).cuda()
out = model.generate(ids, max_new_tokens=1024)

Vision tower and MTP head carry through from base untouched.

Citation

@misc{maldv2026isometryterse,
  title  = {Qwen3.6 27B Isometry Terse},
  author = {Praxis Maldevide},
  year   = {2026},
  note   = {Isometric, norm-preserving merge of Qwen/Qwen3.6-27B with shorter chain-of-thought},
  howpublished = {\url{https://huggingface.co/maldv/Qwen3.6-27B-Isometry-Terse}}
}