Downloads ยท 30 days
294
38% of all-time downloads
sigmanih/Qwen-Qwen3-0.6B-GGUF-Q6_K
Qwen-Qwen3-0.6B-GGUF-Q6_K is a text generation model from sigmanih. Use it when you need the model to write or continue text. The card lists the license as other.
Downloads ยท 30 days
294
38% of all-time downloads
All-time downloads
768
Public
Repo size
623 MB
Likes
1
Public
Click a slice to open those files.
.gguf623 MB ยท 100%
From the Hugging Face model README
โค๏ธ Support & Community: If you find this model helpful, please give this repository a Like on Hugging Face and a โญ Star on our SigmaStudio GitHub!
Qwen-Qwen3-0.6B-GGUF-Q6_K is a production-ready model optimized and published using the Model Hub module of Sigma Studio.
| Specification | Value |
|---|---|
| Model Repository | sigmanih/Qwen-Qwen3-0.6B-GGUF-Q6_K |
| Weight Format | Sconosciuto (sconosciuta) |
| Base Architecture | CausalLM |
| Active Parameters | 0.6B |
| Context Window | 32,768 tokens |
| Total Disk Footprint | 0.0 GB |
| Inference RAM / VRAM | ~1.3 GB VRAM (Full GPU offload) / ~1.1 GB RAM (CPU/Hybrid) |
| Recommended Hardware | GPU with 6-8 GB VRAM (or 16 GB system RAM) |
| Recommended Usage | Edge devices, Real-time voice agents, Mobile & CPU-friendly workloads. |
Evaluated directly on GPU via Sigma Studio Training Lab (Deterministic seed 42, Temp 0.0):
| Benchmark Suite | Score / Accuracy | Total Questions Evaluated | Pass Rate | Test Date | Execution Engine |
|---|---|---|---|---|---|
| Tutti i Benchmark Ufficiali | 46.0% | 46/100 quesiti superati | 46.0% Pass | 2026-09-01 | โก SigmaEngine Direct GPU |
Side-by-side performance comparison between direct zero-overhead answer (No-Thinking) and Chain-of-Thought step-by-step reasoning (Deep Thinking CoT):
โก No-Thinking (Direct Response) : [โโโโโโโโโโโโโโโโโโโโ] 49.1% (28/57 passed)
๐ง Deep Thinking (CoT Reasoning) : [โโโโโโโโโโโโโโโโโโโโ] 41.9% (18/43 passed)
๐ CoT Performance Delta : -7.3%
%%{init: {'theme': 'dark'}}%%
xychart-beta
title "Accuracy (%): No-Thinking vs Deep Thinking"
x-axis ["โก No-Thinking (Direct)", "๐ง Deep Thinking (CoT)"]
y-axis "Accuracy (%)" 0 --> 100
bar [49.1, 41.9]
| Execution Mode | Accuracy (%) | Questions Passed | Operational Profile & Latency |
|---|---|---|---|
| โก No-Thinking (Direct Response) | 49.1% | 28 / 57 | Minimal latency, immediate token-to-first-byte, zero reasoning tokens overhead |
| ๐ง Deep Thinking (CoT Reasoning) | 41.9% | 18 / 43 | Multi-step structured reasoning trace (-7.3%), optimal for math & hard logic |
| Dataset / Benchmark Suite | Domain / Category | Correct / Total | Accuracy (%) | Status |
|---|---|---|---|---|
| ARC-Challenge | Science & Grade-School Reasoning | 5 / 9 | 56% | โก Fair |
| BIG-Bench Hard | Complex Multi-Task Logic & Symbolics | 4 / 7 | 57% | โก Fair |
| GPQA | Graduate-Level Academic Reasoning | 0 / 9 | 0% | โ ๏ธ Low |
| GSM8K | Multi-Step Grade School Math | 8 / 9 | 89% | โ Passed |
| HellaSwag | Commonsense Reasoning & Situational NLI | 4 / 9 | 44% | โก Fair |
| HumanEval | Python Coding (pass@1) | 2 / 7 | 29% | โ ๏ธ Low |
| MATH | Championship Competition Math | 6 / 9 | 67% | โก Fair |
| MBPP | Python Programming with Unit Tests | 5 / 9 | 56% | โก Fair |
| MMLU | General Knowledge & Multi-Subject | 4 / 14 | 29% | โ ๏ธ Low |
| MMLU-Pro | Advanced Multi-Step Reasoning | 0 / 9 | 0% | โ ๏ธ Low |
| TruthfulQA | Factuality & Anti-Hallucination | 8 / 9 | 89% | โ Passed |
| ๐ OVERALL TOTAL | All Evaluated Datasets | 46 / 100 | 46% | ๐ 46% Pass |
Protocol: code_execution, continuation_logprob, cot_generation, letter_logprob ยท temp 0.0 ยท seed 42
Reproducibility hash: SHA256-6F3B83EB3171201A
โ ๏ธ Measured on a slice of the dataset, not the full suite: this score is not comparable with a full-suite run.
Measured on NVIDIA GeForce RTX 5070 Ti โข 15.9 GB VRAM. Two different numbers follow, and they are not interchangeable.
| What was measured | Value | How |
|---|---|---|
| Single-stream decode (what a chat feels) | 389.1 tok/s | one request at a time, on NVIDIA GeForce RTX 5070 Ti |
| Prompt processing | 1812 tok/s | same probe |
| Aggregate throughput during evaluation | 434.8 tok/s | several requests in flight โ not what a single answer runs at |
Speeds on other hardware were not measured and are not guessed here. A single-stream figure from one machine cannot be scaled into a prediction for another: it depends on memory bandwidth, quantization, context length and driver, and the error is large enough to be misleading.
Launch Sigma Studio to enjoy full 1-click GPU hardware acceleration, live monitoring, and visual chat:
# Clone and run Sigma Studio
git clone https://github.com/Sigmanih/SigmaStudio.git
cd SigmaStudio
.\sigma_studio.bat
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model_id = "sigmanih/Qwen-Qwen3-0.6B-GGUF-Q6_K"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat16, device_map="auto")
Qwen-Qwen3-0.6B-GGUF-Q6_K รจ un modello ottimizzato pronto per l'inferenza e l'integrazione locale, pubblicato attraverso ฮฃ-SIGMA Studio.
CausalLM (0.6B parametri)Sconosciuto (sconosciuta)0.0 GB~1.3 GB VRAM (offload GPU completo) | ~1.1 GB RAM (inferenza CPU/ibrida)32,768 tokenTutti i Benchmark Ufficiali46.0% (46/100 quesiti superati)Confronto visuale tra risposta istantanea diretta (No-Thinking) e ragionamento guidato multi-step (Deep Thinking CoT):
โก No-Thinking (Risposta Diretta) : [โโโโโโโโโโโโโโโโโโโโ] 49.1% (28/57 superati)
๐ง Deep Thinking (CoT Reasoning) : [โโโโโโโโโโโโโโโโโโโโ] 41.9% (18/43 superati)
๐ Delta Prestazionale CoT : -7.3%
%%{init: {'theme': 'dark'}}%%
xychart-beta
title "Accuratezza (%): No-Thinking vs Deep Thinking"
x-axis ["โก No-Thinking (Diretto)", "๐ง Deep Thinking (CoT)"]
y-axis "Accuratezza (%)" 0 --> 100
bar [49.1, 41.9]
| Modalitร di Esecuzione | Accuratezza (%) | Quesiti Superati | Profilo Operativo & Latenza |
|---|---|---|---|
| โก No-Thinking (Risposta Diretta) | 49.1% | 28 / 57 | Latenza minima, token-to-first-byte istantaneo, zero overhead di ragionamento |
| ๐ง Deep Thinking (CoT Reasoning) | 41.9% | 18 / 43 | Risoluzione passo-passo multi-step (-7.3%), ideale per logica complessa e matematica |
| Dataset / Suite di Test | Ambito / Dominio | Corretti / Totale | Accuratezza (%) | Esito |
|---|---|---|---|---|
| ARC-Challenge | Ragionamento Scientifico Avanzato | 5 / 9 | 56% | โก Discreto |
| BIG-Bench Hard | Logica Complessa & Compiti Multi-Fase | 4 / 7 | 57% | โก Discreto |
| GPQA | Ragionamento Accademico di Livello Laurea | 0 / 9 | 0% | โ ๏ธ Migliorabile |
| GSM8K | Matematica & Logica Multi-Step | 8 / 9 | 89% | โ Superato |
| HellaSwag | Buon Senso & Comprensione Situazionale | 4 / 9 | 44% | โก Discreto |
| HumanEval | Sintesi Codice Python (pass@1) | 2 / 7 | 29% | โ ๏ธ Migliorabile |
| MATH | Matematica Olimpica & Competitiva | 6 / 9 | 67% | โก Discreto |
| MBPP | Programmazione Python con Unit Test | 5 / 9 | 56% | โก Discreto |
| MMLU | Conoscenza Generale Multidisciplinare | 4 / 14 | 29% | โ ๏ธ Migliorabile |
| MMLU-Pro | Ragionamento Avanzato Multi-Step | 0 / 9 | 0% | โ ๏ธ Migliorabile |
| TruthfulQA | Fattualitร & Resistenza ad Allucinazioni | 8 / 9 | 89% | โ Superato |
| ๐ TOTALE COMPLESSIVO | Tutti i Dataset Valutati | 46 / 100 | 46% | ๐ 46% Pass |
Protocollo: code_execution, continuation_logprob, cot_generation, letter_logprob ยท temp 0.0 ยท seed 42
Impronta di riproducibilitร : SHA256-6F3B83EB3171201A
โ ๏ธ Misurato su una porzione del dataset, non sulla suite intera: il punteggio non รจ confrontabile con uno ottenuto sull'intero.
2026-09-01 su motore deterministico SigmaEngine434.8 tok/s su NVIDIA GeForce RTX 5070 Ti.389.1 tok/s su NVIDIA GeForce RTX 5070 Ti.1812 tok/s.434.8 tok/s โ piu' richieste in volo insieme, non la velocita' di una risposta singola.Se questo modello ti รจ utile o vuoi esplorare l'ecosistema completo:
Creato e distribuito con il Model Hub di ฮฃ-SIGMA Studio (03/11/2026 00:50)