IsValorum/Iris-mini-APEX-I-MiniPlus-V1-GGUF

[!NOTE]

ARCHITECTURE SELECTION GUIDE — MINIPLUS V1 & V2.1 EDITIONS

This repository hosts the MiniPlus V1 edition of Iris-mini. Our releases are precision-engineered for specific hardware budgets and memory topologies. V1 is NOT obsolete or inferior; it represents our leanest, most agile operating profile:

MiniPlus V1 (Lean & Agile Profile): Highly compact footprint with uncompressed F32 router gates, a fully armored Q6_K output head, Q8_0 attention gates, and IQ3_XXS core experts. Both V1 and V2.1 run flawlessly with the vast majority of the model residing in system RAM (DDR4/DDR5), thanks to linear CPU-friendly vectorization that avoids lookup stalls. V1 is dramatically superior to generic community APEX-I-Mini releases (which crush core reasoning down to 2-bit IQ2_S) and flat 3-bit quants.

MiniPlus V2.1 (System RAM Streaming Specialist with Deep Context): Specially prepared to run totally or partially in system RAM (DDR4/DDR5) across deep mathematical and reasoning contexts (up to 256k tokens). Upgrades all 40 shared foundation experts to Q5_K, armors attention gates in Q8_0, and uses linear CPU-friendly vectorization that eliminates AVX2 lookup stalls (+24 to 28+ tok/s). Depending on your processor and memory bandwidth (DDR4/DDR5), streaming generation in system RAM can be almost as fast as having everything in VRAM, while supporting deep context allocating GPU VRAM to the active KV cache while model weights stream from system RAM. All for only ~180 MB more (~13.74 GiB vs ~13.56 GiB)—an overhead that is completely negligible in system RAM.

Which one should you choose? (Official Recommendation: V2.1)

⭐ PRIMARY RECOMMENDATION — Iris-mini APEX-I-MiniPlus V2.1: For virtually all users and deployments, V2.1 is the strictly recommended release. Empirically verified on WikiText-2, V2.1 achieves an outstanding Perplexity of 5.3735 ± 0.1214 (empirically verified on WikiText-2 with near-lossless precision matching Q5_K/Q6_K quality), matching the token fidelity of Q5_K / Q6_K class quantizations while weighing only ~14.7 GB (same footprint as Q3_K_M). Furthermore, it completely eliminates AVX2 CPU stalls, providing blistering +24 to 28+ tok/s streaming under system RAM offload.

MiniPlus V1 Legacy: Maintained for architectural transparency and users seeking specialized configurations for their workflow.

Both editions are handcrafted and vastly outperform flat 3-bit quants and generic community APEX-I-Mini releases. To explore or download the V2.1 edition of Iris-mini, visit: IsValorum/Iris-mini-APEX-I-MiniPlus-V2.1-GGUF

[!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.

Quick Navigation Index

Model Files & Specifications
Comparative Quantization Analysis (vs. Flat Quants & Generic APEX)
Native Multi-Token Prediction (MTP) Co-Pilot
Everyday Laptop Benchmarks (23–26+ tok/s on DDR4)
The 24GB Miracle: Full 256K Context Runs In VRAM!
Hardware Throughput Projections (RTX 30 / 40 / 50)
Why Iris-mini Intentionally Uses Standard APEX (Not V2)
Handcrafted Layer Architecture
Recommended Configuration & Setup

Model Files & Specifications

File Name	File Size	Memory Footprint	BPW	Description
`Iris-mini-MTP.APEX-I-MiniPlus.gguf`	`14.84 GB`	`13.82 GiB`	3.42 BPW	Handcrafted language, math, reasoning & native MTP draft head

Base Architecture: qwen35moe (35B total parameters, approx. 3.2B active per token).
Speculative Decoding: Fully preserved native Multi-Token Prediction head (blk.40).
Target Precision: Armored boundaries (Q3_K), calibrated core experts (IQ3_XXS), 6-bit uncompromised output head (Q6_K).

Comparative Quantization Analysis (vs. Flat Quants & Generic APEX)

Also, don't confuse APEX-I-MiniPlus (Standard) with a generic baseline APEX-I-Mini. Traditional APEX-I-Mini drops core experts aggressively to 2-bit IQ2_S and leaves output.weight at 3-bit Q3_K_M, which creates a noticeable perplexity hit on complex reasoning tasks. Standard MiniPlus avoids that degradation floor while keeping boundary layers in linear Q3_K for single-cycle vectorized AVX2 CPU dequantization (hitting 23 to 26+ tok/s on DDR4 laptops), while protecting output in Q6_K and routers in F32.

To put the numbers in perspective: this cuts nearly 2 GB off a flat 3-bit quant (approx. 15.6 GB), and weighs only about approx. 1 GB more than a generic APEX-I-Mini (approx. 12.5 GB). For that single extra gigabyte of VRAM, you get a massive jump in reasoning and syntactic stability while maximizing CPU/RAM execution throughput.

Take a look at the tensor-by-tensor comparison table below to inspect the exact architectural differences and see why this specific allocation is optimal. That's specifically what this was built for:

Architectural Component	Generic Automated Quants (Flat `Q3_K_S` / `IQ3_S`)	Generic APEX-I-Mini (Baseline Recipe)	Our Handcrafted APEX-I-MiniPlus (Standard / IsValorum)	Perceived Quality & Real-World Impact
Output Head (`output.weight`)	Flat `IQ3_S` / `Q3_K_S` (approx. 3.44 BPW)	Inherits base type `Q3_K_M` (approx. 3.44 BPW unarmored)	`Q6_K` (approx. 6.56 BPW uncompromised)	Eliminates Syntax & Vocabulary Hallucinations: Low-bit output heads cause tokenizer classification noise, breaking code indentation, brackets (`{}`, `[]`), math symbols, and domain terms. `Q6_K` preserves near-FP16 output classification.
Expert Routers (`ffn_gate_inp.weight`)	Blindly quantized to 3-bit / unoptimized	Inherits base type `Q3_K_M` (approx. 3.44 BPW compressed)	`F32` uncompressed (32.0 BPW, 2 MB/layer)	Zero Router Drift: In micro-expert models, even minuscule quantization errors in router logits misdirect tokens to wrong experts. Retaining uncompressed `F32` guarantees 100% routing fidelity with virtually zero memory overhead (approx. 80 MB total).
Attention & Language (`attn_output`, `attn_qkv`)	Flat `IQ3_S` / `Q3_K_S`	`Q3_K` on 34 middle layers (L3–36), `Q4_K` on 6 edge layers	`Q6_K` for `attn_output`, `Q3_K` / `Q4_K` + `imatrix`	Contextual Precision & CPU Throughput: Combines uncompromised `Q6_K` for the output projection with fast vectorized linear blocks for attention, balancing retrieval accuracy with maximum token streaming speed on CPU/RAM.
Attention Gates (`attn_gate.weight`)	Blindly compressed to 3-bit	Compressed to `Q3_K` (middle) / `Q4_K` (edges)	`Q4_K` / `Q8_0` (linear high-precision)	Attention Routing Dynamics: High-precision linear gating modulating query-key projections without CPU dequantization latency.
*Shared Foundation Expert (`ffn__shexp`)**	Flat `IQ3_S` / `Q3_K_S` (3.44 BPW)	Linear `Q4_K` (middle) / `Q5_K` (edges)	Linear `Q4_K` (middle) / `Q5_K` (edges) + `imatrix`	Foundational Knowledge Stability: Keeps the universal pathway in high-fidelity linear blocks, eliminating quantization drift while maintaining rapid single-cycle dequantization.
Core MoE Layers (Middle: 10–29)	Flat `IQ3_S` / `Q3_K_S` (uniform bit-rate across all layers)	Aggressive `IQ2_S` (2.50 BPW)	`IQ3_XXS` (3.06 BPW) + calibrated `imatrix`	Above the Quality Threshold: Generic 2-bit `IQ2_S` baselines drop below the critical quality floor for 35B MoEs, resulting in perplexity spikes on reasoning tasks. Our `IQ3_XXS` with imatrix achieves deep compression (272 MiB → 98 MiB per block) without sacrificing logic.
Edge MoE Layers (Layers 0–9 & 30–39)	Flat `IQ3_S` / `Q3_K_S` (no layer-wise gradient)	`Q3_K` (limited to first/last 5 layers only: L0–4, L35–39)	`Q3_K` (expanded to 10 input & 10 output layers)	AVX2 Single-Cycle Speed: Expanded 10+10 layer protection using linear `Q3_K` blocks enables single-cycle vectorized AVX2 CPU dequantization, unlocking 23 to 26+ tok/s on budget DDR4 laptops.
MTP Draft Block (`blk.40`)	Stripped with `--no-mtp` or broken	Crushed to `IQ2_S` / tier precision	Preserved in `IQ3_S` / `Q3_K` & `IQ4_NL`	Speculative Decoding Speedup: Maintains 58%–65% candidate acceptance rate, yielding 1.6x–1.75x real-world token speedup without speculative rejection waste.
Normalization & Biases	Often degraded	Standard	`F32` uncompressed	Numerical Stability: Prevents cumulative floating-point underflow/overflow across deep 40-layer computation.

Native Multi-Token Prediction (MTP) Co-Pilot

Most automated community releases strip or break the native Multi-Token Prediction head using --no-mtp. APEX-I-MiniPlus fully preserves and calibrates the native prediction block (blk.40):

Zero-Cost Speculative Acceleration: Unlike external draft models that consume separate VRAM and memory bandwidth, Iris-mini's native MTP head is integrated directly into the weights.
Empirical Acceptance Rate: 58.8% to 65.5% of predicted candidate tokens are accepted on full GPU offload.
Token Yield: Delivers 1.60 to 1.75 tokens per forward step on standard text, peaking at 2.0+ tokens/step during continuous code and prose generation.
Net Speedup: Provides approx. 1.6x faster real-world generation without quality degradation.

Everyday Laptop Benchmarks (23–26+ tok/s on DDR4)

VRAM Allocation: 3.8 GB VRAM utilized on budget 4GB/6GB GPUs.
System Memory: 32GB DDR4 holds the remaining layers.
Prefill Speed: 300 to 410 tokens/second sustained across dense inputs.
Generation Speed: 23 to 26+ tokens/second sustained on standard DDR4 RAM!

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

Context Length	Model Weights (Est.)	KV Cache (q8_0, 4 slots)	Compute Buffers	Total GPU VRAM (Est.)	Hardware Feasibility
32,512 (32k)	`13.82 GiB`	`0.58 GiB`	`1.80 GiB`	`16.20 GiB`	Full offload on 24GB; partial on 16GB
64,512 (64k)	`13.82 GiB`	`0.92 GiB`	`1.95 GiB`	`16.69 GiB`	Effortless fit on 24GB GPUs
128,640 (128k)	`13.82 GiB`	`1.58 GiB`	`2.22 GiB`	`17.62 GiB`	Effortless fit on 24GB GPUs
262,144 (Full 256K)	`13.82 GiB`	`2.92 GiB`	`2.80 GiB`	`19.54 GiB`	FULL 256K NATIVE IN VRAM!

Hardware Throughput Projections (RTX 30 / 40 / 50)