CALIBRATED 2026-07-27 · REC 003
Local AI Frontier
BM-003TESTED 2026-07-27

Qwen3-Coder-30B-A3B quantization sweep: Q4_K_M vs. Q6_K vs. Q8_0 on Arc B60 Pro

On 24GB Arc, Q6_K is the sweet spot: quality recovers to within 2 points of Q8_0 on HumanEval while staying 38% faster and fitting comfortably. Q8_0 barely fits and leaves no room for context. Q4_K_M is the budget pick when VRAM is tight.

vulkanq6-kllama.cpp b3500codingintel-arcvulkanquantization30b-class

Configuration

Model
Qwen3-Coder-30B-A3B-Instruct
Artifact
Qwen3-Coder-30B-A3B-Instruct (multiple GGUF quants)
Checksum
pending
Quantization
q6-k
Context length
8,192 tokens
Backend
vulkan
Runtime
llama.cpp b3500
Settings
-ngl 99 -c 8192 -t 8 -b 512
Workload pack
workload-pack-v1
Author
Edgar
Reviewer
Edgar

Results

Identical workload pack run three times, one per quantization.

38.6tok/s

GEN · Jitori PC

412tok/s

PROMPT

740ms

TTFT

18.1GB

VRAM USED

MACHINEPROMPT tok/sGEN tok/sTTFT msVRAM GBRAM GBPOWER W
Jitori PC
Q6_K quant. Fits with ~1.4GB headroom.
398.127.498022.63.4250
Jitori PC
Q8_0 quant. Saturates VRAM; no context headroom.
380.519.8132024.04.1252
Jitori PC
Q4_K_M quant. From bm-001, included for direct comparison.
412.338.674018.13.2245

Quality scores

CODING

0.81/ 0-1

HumanEval pass@1 (Q4_K_M)

CODING

0.84/ 0-1

HumanEval pass@1 (Q6_K)

CODING

0.86/ 0-1

HumanEval pass@1 (Q8_0)

Limitations

  • Quality deltas are within typical run-to-run variance for single-pass HumanEval; treat as directional, not definitive.
  • VRAM figures include KV cache at -c 8192; longer contexts shift the Q6_K vs Q8_0 boundary.
  • Single machine (Jitori) — backend effects on other Arc SKUs may differ.
NO COMMERCIAL INTEREST IN THIS RESULT