Qwen3.8-27B Base VAL-APEX-I NanoPlus V3 GGUF

Independently computed using dedicated cloud compute clusters, funded 100% out-of-pocket.

If you consider these releases to be of high quality and they help you in your day-to-day workflow, consider supporting on Ko-fi. Every contribution directly sustains future cloud compute hours and helps support my university studies.

Original Qwen checkpoint · Revised V3 tensor allocation · Native 256K context

Official VAL-APEX-I NanoPlus V3 quantization of Qwen/Qwen3.8-27B. This edition quantizes the original Qwen checkpoint without applying Huihui abliteration or EfficientThink fine-tuning.

Integrated Multi-Token Prediction (MTP): This GGUF includes one MTP block (blk.64.*, 15 tensors) in the same file as the main language model. Its eight matrices use Q4_0 ×6, Q4_1 ×1 and Q5_K ×1; its seven normalization tensors remain F32. MTP speculative decoding depends on support and configuration in the inference backend.

VAL-APEX-I stands for: Vector-calibrated Asymmetric Layer-wise Outlier-preserving Recurrent-aware Unified Matrix-quantization

[!NOTE]

EXPLORE THE QWEN3.8 27B VAL-APEX-I EDITIONS

Choose a model variant and quantization profile:

• Qwen3.8 27B — MiniPlus V3 — 14.16 GB.
• Swift-1.5 Qwen3.8 27B Uncensored MTP — MiniPlus V3 — 14.16 GB.
• Swift-1.5 Qwen3.8 27B Uncensored MTP — NanoPlus V3 — 12.18 GB. Collections: MiniPlus Abliterated 27B · MiniPlus Standard 27B · NanoPlus Abliterated 27B · NanoPlus Standard 27B

[!IMPORTANT]

COMPACT 12.18 GB BUILD

This V3 release uses a revised per-tensor allocation with an actual 12.18 GB file size and 3.57 whole-file BPW. Its reported WikiText-2 PPL is 6.2577 ± 0.5011, or +0.4635 (+8.00%) relative to the measured unmodified BF16 Base reference. The audit below records every tensor family, the block-0 exceptions and the integrated MTP head.

---

Quick Navigation Index

1. Quantization Metrics & Tensor Map
2. Independent Community Benchmarks (NanoBench v0.4.0)
3. Model Files & Technical Specifications
4. Native Context & Runtime Memory
5. Recommended Configuration & Setup
6. Recommended Generation Parameters
7. Native Chat Template & Reasoning
8. Community Compute Fund & Student Support

---

1. Quantization Metrics

Size & Quality Metrics

[!NOTE] Independent Hardware & Compute Transparency: Generating multi-million token calibration matrices and running certified perplexity evaluations on local architectures requires dedicated high-end cloud nodes funded 100% out-of-pocket. If you rely on these MiniPlus/NanoPlus builds for your local workloads, contributions to the IsValorum Compute Fund on Ko-fi directly sponsor the GPU hours required for future Day-0 handcrafted quantizations.

Independent Community Benchmarks (NanoBench v0.4.0)

Independent multi-harness empirical evaluation conducted by community member @userr99 on a dedicated local workstation (24 GB VRAM / 64 GB RAM using llama.cpp). Evaluated across the 12-task autonomous coding and agentic suite on NanoBench v0.4.0 (public discussion thread & test logs):

Key Architectural Insights:

• Zero Tool-Calling Degradation: Achieved a unanimous 100/100 in Agentic tasks across all three harnesses, confirming that protecting output.weight in Q5_K and DeltaNet recurrent memory states (ssm_a in F32, ssm_beta in Q8_0) prevents JSON schema rejections and state drift over multi-step workflows.
• Universal Harness Compatibility: Surpassed the 90+ score barrier across all three test harnesses without entering infinite retry loops or syntax divergence.

Tensor Precision Map

x denotes tensor counts. All columns reflect exact GGUF tensor headers verified in tensor-map.tsv. Normalizations exclude separately listed SSM norms.

Both Swift/Base/Huihui files contain 866 audited tensors: 851 main-model tensors plus 15 MTP tensors. The main model is a dense hybrid architecture with 64 layers: 48 Gated DeltaNet linear-attention layers and 16 full-attention layers. The additional blk.64.* block is the integrated MTP head.

V3 architectural allocation: NanoPlus V3 features an elevated 3-bit floor across all SwiGLU MLP blocks (ffn_gate and ffn_up set uniformly to IQ3_XXS across all 64 layers; ffn_down uses IQ4_XS in outer layers 0-7 / 56-63 and IQ3_XXS in core layers 8-55). Output head (output.weight) is calibrated to Q5_K (5.50 bpw) to maximize token throughput and preserve dynamic range, entirely eliminating the linguistic decay in non-English contexts associated with 2-bit sub-allocations.

Actual layer-0 formats: blk.0.attn_gate.weight, blk.0.attn_qkv.weight, blk.0.ssm_beta.weight and blk.0.ssm_out.weight use Q4_0. All 48 ssm_alpha tensors remain F32 for mathematical stability.

Integrated MTP: blk.64.* contains 424,699,392 stored parameters in 15 tensors: seven F32 tensors and eight matrices (Q4_0 x6, Q4_1 x1, Q5_K x1). Runtime acceleration depends on backend support and configuration; speculative decoding operates via native acceptance drafting.

Complete tensor-by-tensor audit: names, formats, shapes and parameter counts.

---

2. Model Files & Technical Specifications

• Source: Qwen/Qwen3.8-27B.
• Architecture: qwen35; 64 main layers (48 Gated DeltaNet + 16 full-attention) plus one MTP block.
• Stored parameters: 27,320,697,856, including MTP tensors.
• Tensor count: 866, including the 15-tensor MTP head.
• Native context: 262,144 tokens.
• Calibration: 1,000 chunks × 4,096 tokens (4,096,000 tokens), as reported by the author. The GGUF records 1,000 calibration chunks and 496 importance-matrix entries.
• Scope: This repository currently contains the language GGUF. No companion vision projector is included.

Whole-file BPW is 8 × file bytes / stored parameter count, including the file's metadata and MTP head. It is not a uniform tensor quantization type.

---

3. Native Context & Runtime Memory

The model's native context is 262,144 tokens. File size describes the stored weights; runtime memory also includes the KV cache, recurrent state, compute buffers, batch settings and any optional components. Choose context and offload settings according to the allocations reported by your runtime.

---

4. Recommended Configuration & Setup

llama.cpp Server

llama-server \
  -m Qwen3.8-27B-VAL-APEX-I-NanoPlus-V3.gguf \
  --jinja \
  -ngl 99 \
  -c 8192 \
  --host 127.0.0.1 \
  --port 8080

The command requests GPU offload; adjust offload and context to the available memory. In LM Studio, load the GGUF and retain its embedded Qwen chat template. The integrated MTP tensors are listed in the audit; speculative decoding requires explicit support and configuration in the chosen backend.

---

5. Recommended Generation Parameters

These are the upstream Qwen defaults, distinct from the WikiText-2 evaluation settings.

---

6. Native Chat Template & Reasoning

The native Qwen Jinja chat template is embedded in this GGUF. Use --jinja where supported and preserve the native formatting for messages, tool calls and thinking blocks. Reasoning controls depend on the backend's implementation of the template arguments.

---

8. Community Compute Fund & Student Support

All IsValorum quantizations will always remain completely free and open to the public without paywalls.

However, cloud GPU compute is expensive and funded entirely out-of-pocket. If you find these handcrafted builds to be of high quality and they benefit your day-to-day workflow, you can support future Day-0 releases on Ko-fi. Every contribution directly funds the cloud compute hours required for new model runs and helps support my university education:

(When supporting on Ko-fi, feel free to leave a note with your Hugging Face handle and the specific model you would like prioritized).