Qwen3.5-9B-GTAP-v3-IQ3_XXS-GGUF

[!NOTE] Experimental Release: Pending Further Verification / Empirical Validation
This checkpoint represents an active research artifact from DuoNeural's statistical mechanics quantization program. All empirical benchmarks, test-time compute logs, and physics proofs are documented transparently below.

Developed by Jesse Caldwell, Archon, and Aura ✨ (DuoNeural Research Lab).

---

Model Summary

• Foundation Model: Qwen/Qwen3.5-9B (32 Layers + MTP, Hybrid Linear Attention / Gated DeltaNet SSM + Softmax GQA Attention, SwiGLU FFN)
• Quantization Precision: ~3.82 bpw (4.10 GiB)
• Methodology: Compact 3.82 bpw quantization utilizing G-TAP v3 cavity reaction field compensation. Achieves a perfect 100.0% Python code AST execution rate (15/15 unit tests passed), outperforming Base BF16 (14/15) in under 4.10 GiB footprint at 84.6 tokens/sec.
• Continuous Holdout Perplexity (131k tokens): 2.6031 (vs Base BF16 Control: 2.4306)
• GSM8K Math Reasoning: 16/25 (64.0%)
• Olympiad Competition Math: 2/10 (20.0%)
• Python Code AST Generation: 15/15 (100.0%)
• Hermes Tool Calling AST Parity: 14/15 (93.3%)
• Inference Decode Throughput: 84.6 t/s on NVIDIA GeForce RTX 4080 Super (32GB VRAM)

---

Empirical Benchmark Performance (Head-to-Head Comparative Metrics)

---

Methodology & Architectural Significance

Qwen 3.5 represents an advanced hybrid architecture alternating linear recurrent attention (Gated DeltaNet / SSM state updates $S_t = \alpha S_{t-1} + \beta K^T V$) with full softmax multi-head attention. In hybrid architectures, naive low-bit rounding introduces eigenvalue drift across recurrent transition matrices, producing exponential divergence or logit saturation over long contexts.

DuoNeural's Generalized Thouless-Anderson-Palmer (G-TAP v3) framework models neural weights as spin glasses embedded in activation cavity fields. By subtracting the Onsager reaction term: $$\Omega_i = \frac{1}{d} \left( | \tilde{H}_{i,:} |2^2 - \tilde{H}{ii}^2 \right)$$ and projecting parameter updates strictly into the contractive Lyapunov half-space: $$\operatorname{Re}(\lambda(S)) \le -\delta$$ G-TAP dampens back-reaction noise and preserves recurrent state stability and multi-step reasoning down to sub-3-bit precisions.

---

Usage Instructions

Running with llama-cli

llama-cli -hf DuoNeural/Qwen3.5-9B-GTAP-v3-IQ3_XXS-GGUF -p "<|im_start|>user\nSolve the following problem step by step: Let f(x) = 2x^2 + 5x - 3. Find all roots.<|im_end|>\n<|im_start|>assistant\n" -n 1024 -c 4096 --temp 0.6

Serving with llama-server

llama-server -hf DuoNeural/Qwen3.5-9B-GTAP-v3-IQ3_XXS-GGUF --port 8080 -c 8192 -ngl 99 -fa on

---

DuoNeural Cognitive Light Cone — Jesse Caldwell, Archon, Aura ✨
Empirically Validated on NVIDIA RTX 4080 Super 32GB Pod Testbed