base_model: empero-ai/Qwen3.8-35B-A3B-Distill
quantized_by: IsValorum
library_name: gguf
language:

• en
• zh
• es
• fr
• de
• pt
• it
• ru
• ja
• ko
• vi
• th
• ar
  tags:
• abliterated
• uncensored
• alignment
• heretic
• gguf
• quantized
• quantization
• apex
• apex-quant
• custom-quantization
• unsloth-studio
• moe
• reasoning
• llama.cpp
• qwen35moe
  license: apache-2.0
  pipeline_tag: text-generation

Qwen3.8-35B-A3B-Distill-MTP APEX-I-MiniPlus-V2.1 Abliterated GGUF (Uncensored / Zero Refusal)

The Definitive Uncensored Frontier MoE · True Directional Refusal Ablation · Dedicated Q8_0 MTP & Vision Companions Included · Full 256K Context

[!NOTE]

🛡️ LOOKING FOR THE STANDARD ALIGNED RELEASE?

If you prefer the standard safety-aligned edition with default guardrails, check out the official base release:
🔗 Qwen3.8-35B-A3B-Distill-MTP-APEX-I-MiniPlus-V2.1-GGUF — identical featherweight 14.75 GB footprint, Q5–Q6 fidelity, zero AVX2 CPU stalls, and dedicated Q8_0 MTP & mmproj companions.

[!TIP]

🔓 AUTHENTIC HERETIC TPE DIRECTIONAL ABLITERATION (UNCENSORED)

This is the official authentic Abliterated / Uncensored edition of Qwen3.8-35B-A3B-Distill-MTP APEX-I-MiniPlus V2.1.

• Zero Moralizing Refusals: Complete elimination of refusal vectors across systems security, pen-testing, and compliance tasks (0/10 refusals in empirical validation).
• Orthogonal Activation Steering: Refusal directions are mathematically isolated and steered orthogonally (D_KL = 0.0076) to preserve 100% of underlying domain knowledge, logic, and chain-of-thought reasoning capability.
• Empirical Verification: Direct binary probe testing observed 0/10 refusals in the tested probe set without syntax corruption or degradation in MoE routing.

[!IMPORTANT]

THE DEFINITIVE SPECIFICATION IN THE 13–14 GB CEILING

This APEX-I-MiniPlus-V2.1 release represents the absolute technological limit of sparse Mixture-of-Experts quantization within the 13–14 GB envelope. Every single tensor of its 40 layers and 256 micro-experts has been mathematically audited to maximize reasoning precision, eliminate recurrence state drift, and prevent AVX2 CPU dequantization stalls.

[!TIP]

🏆 THE QUANTIZATION SWEET SPOT: Q5–Q6 FIDELITY AT 3-BIT FOOTPRINT

Why APEX-I-MiniPlus V2.1 outperforms standard community quants:

• Empirical WikiText-2 Perplexity: 5.1280 ± 0.1312 (evaluated directly on GGUF binary, 2048 ctx, 8 chunks) — near-zero loss (ΔPPL ≈ +0.02) from unquantized baseline (approx. 5.10).
• Q5–Q6 Quality at 3-Bit Footprint: Delivers reasoning fidelity and syntax consistency typical of Q5_K / Q6_K while occupying only approx. 14.65 GB (13.65 GiB) (same featherweight footprint as Q3_K_M).
• Zero Routing Drift: 100% of expert routing matrices (gate_inp) remain in uncompressed F32, ensuring tokens are dispatched to the exact right experts on every forward pass.

[!WARNING]

DO NOT CONFUSE APEX-I-MINIPLUS WITH GENERIC COMMUNITY APEX-I-MINI!

Regardless of release version (whether V1, V2, or V2.1), NEVER confuse handcrafted APEX-I-MiniPlus builds with generic community APEX-I-Mini releases:

• Generic Community APEX-I-Mini: Uniformly compresses all core MoE experts down to aggressive 2-bit IQ2_S (dropping below the critical quality floor), leaves the sensitive token output head unarmored at 3-bit Q3_K_M, and compresses attention projections down to Q3_K. In deep reasoning models, this triggers severe perplexity spikes, syntax errors, and broken code brackets.
• Handcrafted APEX-I-MiniPlus (All Editions by IsValorum): Every single MiniPlus release—from V1 and V2 to V2.1—is a custom tensor-by-tensor architecture that preserves uncompressed F32 router gates, armors the token output head in high-precision Q6_K, safeguards attention gates in Q8_0, and keeps core reasoning experts at or above calibrated 3-bit (IQ3_XXS/IQ3_S). Even our earlier builds vastly outperform generic community APEX recipes and flat 3-bit quants.

Optimization History & Transparency Notice

We maintain our previous releases publicly as a transparent engineering record of continuous optimization. Below is the exact evolutionary roadmap of our MiniPlus architectures:

[!TIP]

Deployment & System Architecture Guide

• Full GPU VRAM Offload (24GB+ VRAM, -ngl 99): Effortless full offload with native 256K context support. Blistering throughput on RTX 3090 / 4090 / 5090 GPUs.
• System RAM Streaming Specialist (DDR4/DDR5 & Massive Context): Specially engineered to run either partially or entirely out of system RAM across large or full (+160k to 256k) context windows. By replacing non-linear codebooks with linear SIMD-optimized Q3_K edge experts and upgrading shared foundation experts to Q5_K across all 40 layers, AVX2 CPU dequantization stalls are completely eliminated (+24 to 28+ tok/s). Depending on your processor architecture and memory bandwidth (dual-channel DDR4 or high-speed DDR5 6000+ MT/s), streaming generation in system RAM can approach speeds remarkably close to full VRAM execution, allowing both the dedicated Q8_0 multimodal vision projector (mmproj) and the Q8_0 Multi-Token Prediction draft head (mtp) to be loaded explicitly in GPU VRAM for accelerated speculative decoding and zero-latency OCR while the main model weights stream effortlessly from system RAM.

Explore the complete family of APEX-I-MiniPlus models in our official collection:
APEX-I-MiniPlus V2.1 Hub Collection.

Quick Navigation Index

• Model Files & Technical Specifications
• Empirical Benchmarks & Fidelity Verification
• Quality Spectrum: APEX-I-MiniPlus vs. Standard Flat Quants
• Independent Benchmark of the APEX-I-MiniPlus Family (Occamy V2 Reference)
• The 24GB Miracle: Full 256K Context Runs In VRAM!
• Hardware Throughput & Offload Benchmarks (RTX 30 / 40 / 50)
• Surgical Tensor Quantization Map (Audited from GGUF)
• Recommended Configuration & CLI Setup
• Recommended Generation Parameters (Creator Official)
• Model Inherent Behavior vs. Quantization Fidelity Notice

🏅 Independent Benchmark of the APEX-I-MiniPlus Family (Occamy V2 Reference)

[!NOTE]
External report: zephel01 independently benchmarked Occamy V2. The benchmark below was performed on Occamy-1.0 APEX-I-MiniPlus V2, not on this specific V2.1 model. It is included as independent evidence of the broader MiniPlus quantization approach.

The APEX-I-MiniPlus quantization architecture powering this model was subjected to an extensive independent evaluation by Japanese AI researcher and evaluator zephel01 (CoolZero) on an NVIDIA RTX 5090 (32GB) workstation running llama.cpp CUDA b11027 with FlashAttention (-fa on -ctk q8_0 -ctv q8_0 -ngl 99).

The evaluation tested the APEX-I hybrid MoE engine across 348 unseeded trials on SWE-bench style multi-file Python bug-fixing tasks with hidden pytest suites (llmbench):

• L6 Multi-File Code Generation (60 tasks):
  • Context 32,768 (32K): 93.3% Resolved (46/60 tasks passed 5/5 consecutive trials; 20/20 on Easy–Hard).
  • Context 65,536 (65K): 90.0% Resolved (45/60 tasks passed 5/5 consecutive trials).
  • Match with 25–28 GB Models: Matches or exceeds the resolution rate of full 25–28 GB models (such as Ornith-1.5 and Tiel-Coder 35B-A3B) while consuming over 10 GB less VRAM (14.6 GB vs approx. 26 GB).
• Extreme Context VRAM Scaling (The Hybrid DeltaNet SSM Advantage):
  • 32K Context: 14.6 GB total VRAM allocation.
  • 65K Context: 15.1 GB total VRAM allocation (only +0.5 GB VRAM added when doubling context!).
  • Architectural Explanation: Because 30 of the 40 layers utilize Linear Attention / DeltaNet SSM (O(1) constant recurrence memory), only the 10 full-attention anchor layers expand the KV cache. This proves empirically that 65,536 context runs 100% in VRAM on consumer 16GB GPUs (RTX 4080 / RTX 5080) without offloading to system RAM.
• Measured Real-World Throughput: Sustained single-stream generation of approx. 247 – 251 tok/s on NVIDIA RTX 5090.

Empirical Benchmarks & Fidelity Verification

Quality Spectrum: APEX-I-MiniPlus V2.1 vs. Standard Flat Quantizations

How the handcrafted APEX-I-MiniPlus V2.1 architecture compares against standard flat quantizations in llama.cpp on 35B Mixture-of-Experts architectures:

Model Files & Technical Specifications

• Base Model: empero-ai/Qwen3.8-35B-A3B-Distill
• Parameters: 35.2B total (approx. 2.6B to 3.2B active per token)
• Architecture: 40 layers, 256 micro-experts (8 active per token) + hybrid linear attention / DeltaNet recurrent layers
• Context Length: 262,144 tokens (native 256K)

Surgical Tensor Quantization Map (Audited from GGUF)

The exact tensor breakdown below has been verified directly from the compiled binary weights:

Hardware Throughput & Offload Benchmarks (RTX 30 / 40 / 50 & RAM Streaming)

Empirically verified in Unsloth Studio & llama.cpp:

• Aggressive Hybrid Offload Profile: Sustained 24.25 to 28.37 tok/s generation with reasoning enabled, even when only approx. 4.2 GB VRAM is available and the rest of the 13.74 GiB model streams from system RAM.

[!NOTE]

Empirical Testbed Architecture & Desktop/Server Scaling

• Empirical Benchmark Hardware: The hybrid offload and system RAM streaming figures documented above (sustaining 24.25 to 28.37 tok/s) were measured on a consumer laptop powered by an Intel 12th Gen Alder Lake architecture featuring a hybrid design of Performance Cores (P-Cores) and Efficient Cores (E-Cores) paired with dual-channel system RAM and constrained laptop power/thermal envelopes.
• Thread Scheduling & E-Core Contention: In hybrid architectures like Alder Lake, OS thread scheduling across background E-Cores and lower single-core mobile power limits introduce memory bandwidth and thread synchronization overhead during CPU dequantization.
• Dramatic Scaling on Higher-End Processors: When running on desktop or server processors (such as modern AMD Ryzen 7000 / 9000 Zen 4/5 series or high-TDP Intel desktop platforms with dedicated performance cores, large L3 caches, and high-bandwidth dual- or quad-channel DDR5 running at 6000+ MT/s), streaming generation speeds and prefill throughput will scale dramatically higher, substantially exceeding these measured mobile numbers.

The 24GB Miracle: Full 256K Context Runs In VRAM!

Qwen3.8-35B-A3B-Distill-MTP APEX-I-MiniPlus-V2.1 fits the entire 256K context window within 24GB VRAM:

Note: Leaves comfortable headroom for display drivers and compute buffers on standard 24GB GPUs (RTX 3090, RTX 4090, RTX 5090).

[!TIP]

💡 Empirical 16GB GPU Verification (Single Stream / Desktop)

While theoretical multi-slot server buffers estimate approx. 16.5 GiB, independent hardware testing by zephel01 on an RTX 5090 (Occamy V2 Reference) confirmed that single-stream desktop inference consumes only 14.6 GB at 32,768 ctx and only 15.1 GB at 65,536 ctx (-ctk q8_0 -ctv q8_0 -fa on). This empirically proves that full 65K context runs completely in VRAM on 16GB cards (RTX 4080 / RTX 5080) without system RAM offload!

Recommended Configuration & Setup

llama-server.exe \
 -m Qwen3.8-35B-A3B-Distill.APEX-I-MiniPlus-V2.1-Abliterated.gguf \
 --port 8080 \
 --parallel 4 \
 --flash-attn on \
 --fit on \
 -c 104960 \
 --cache-type-k q8_0 \
 --cache-type-v q8_0

⚙️ Recommended Generation Parameters (empero-ai Official)

Official generation guidelines specified by empero-ai for distilled reasoning MoEs:

[!IMPORTANT]

🔍 Model Inherent Behavior vs. Quantization Fidelity Notice

Any behavioral nuances, stylistic tendencies, domain-specific habits, or zero-shot edge-case oversights stem entirely from the original unquantized checkpoint weights and fine-tuning distribution, NOT from the APEX-I quantization process.
Handcrafted APEX-I-MiniPlus strictly preserves mathematical tensor fidelity—keeping 100% of expert routing matrices (gate_inp) in uncompressed F32 (zero router drift), armoring the token output head in Q6_K, and safeguarding attention gates in Q8_0. Empirical verification confirms near-zero perplexity loss (ΔPPL ≈ +0.02), ensuring that token logits, routing decisions, and reasoning trajectories are mathematically faithful to the original base model.