---

language:

• en
• it

license: other

base_model:

• ornith-ai/Ornith-1.0-35B

tags:

• text-generation
• sigma-studio
• sigmanih
• conversational
• custom-model
• gguf
• llama.cpp
• quantized
• q4_k_m

pipeline_tag: text-generation

---

<div align="center">

⚡ ornith-ai-Ornith-1.0-35B-GGUF-Q4_K_M

High-Performance Model Published via Σ-SIGMA Studio

   

</div>

> ❤️ Support & Community: If you find this model helpful, please give this repository a Like on Hugging Face and a ⭐ Star on our SigmaStudio GitHub!

🌐 English Overview

ornith-ai-Ornith-1.0-35B-GGUF-Q4_K_M is a production-ready model optimized and published using the Model Hub module of Sigma Studio.

⚙️ Technical Specifications & Architecture

| Specification | Value |

| :--- | :--- |

| Model Repository | sigmanih/ornith-ai-Ornith-1.0-35B-GGUF-Q4_K_M |

| Weight Format | GGUF (Q4_K_M) |

| Base Architecture | qwen35moe |

| Active Parameters | 35B |

| Context Window | 262,144 tokens |

| Transformer Layers | 40 |

| Hidden Dimension | 2048 |

| Total Disk Footprint | 19.71 GB |

| Inference RAM / VRAM | ~20.8 GB VRAM (Full GPU offload) / ~20.3 GB RAM (CPU/Hybrid) |

| Recommended Hardware | GPU with 24 GB VRAM (e.g. RTX 3090 / 4090 or 32-64 GB RAM) |

| Recommended Usage | Flagship frontier intelligence, Deep research & multi-step mathematics. |

🏆 Official Benchmark Performance

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 | 72.0% | 72/100 quesiti superati | 72.0% Pass | 2026-09-08 | ⚡ SigmaEngine Direct GPU |

🧠 Reasoning Mode Comparison: No-Thinking vs Thinking

Side-by-side performance comparison between direct zero-overhead answer (No-Thinking) and Chain-of-Thought step-by-step reasoning (Deep Thinking CoT):

⚡ No-Thinking (Direct Response) : [████████████████░░░░] 80.7%  (46/57 passed)
🧠 Deep Thinking (CoT Reasoning) : [████████████░░░░░░░░] 60.5%  (26/43 passed)
📈 CoT Performance Delta        : -20.2%

%%{init: {'theme': 'dark'}}%%
xychart-beta
    title "Accuracy (%): No-Thinking vs Deep Thinking"
    x-axis ["⚡ No-Thinking (Direct)", "🧠 Deep Thinking (CoT)"]
    y-axis "Accuracy (%)" 0 --> 100
    bar [80.7, 60.5]

| Execution Mode | Accuracy (%) | Questions Passed | Operational Profile & Latency |

| :--- | :---: | :---: | :--- |

| ⚡ No-Thinking (Direct Response) | 80.7% | 46 / 57 | Minimal latency, immediate token-to-first-byte, zero reasoning tokens overhead |

| 🧠 Deep Thinking (CoT Reasoning) | 60.5% | 26 / 43 | Multi-step structured reasoning trace (-20.2%), optimal for math & hard logic |

📋 Per-Dataset Evaluation Breakdown

| Dataset / Benchmark Suite | Domain / Category | Correct / Total | Accuracy (%) | Status |

| :--- | :--- | :---: | :---: | :---: |

| ARC-Challenge | Science & Grade-School Reasoning | 9 / 9 | 100% | ✅ Passed |

| BIG-Bench Hard | Complex Multi-Task Logic & Symbolics | 5 / 7 | 71% | ✅ Passed |

| GPQA | Graduate-Level Academic Reasoning | 2 / 9 | 22% | ⚠️ Low |

| GSM8K | Multi-Step Grade School Math | 8 / 9 | 89% | ✅ Passed |

| HellaSwag | Commonsense Reasoning & Situational NLI | 6 / 9 | 67% | ⚡ Fair |

| HumanEval | Python Coding (pass@1) | 7 / 7 | 100% | ✅ Passed |

| MATH | Championship Competition Math | 5 / 9 | 56% | ⚡ Fair |

| MBPP | Python Programming with Unit Tests | 7 / 9 | 78% | ✅ Passed |

| MMLU | General Knowledge & Multi-Subject | 9 / 14 | 64% | ⚡ Fair |

| MMLU-Pro | Advanced Multi-Step Reasoning | 6 / 9 | 67% | ⚡ Fair |

| TruthfulQA | Factuality & Anti-Hallucination | 8 / 9 | 89% | ✅ Passed |

| 🏆 OVERALL TOTAL | All Evaluated Datasets | 72 / 100 | 72% | 🏆 72% Pass |

Protocol: code_execution, continuation_logprob, cot_generation, letter_logprob · temp 0.0 · seed 42

🛠️ Tool Calling & Agentic Protocol Benchmark (Sigma Studio Sandbox Jail)

Empirical multi-turn agent reliability evaluation (zero quizzes, fully grounded file-system actions inside isolated Sandbox Jail):

| Protocol Metric | Outcome | Validation Criteria & Details |

| :--- | :---: | :--- |

| Tool Protocol Adherence | 100.0% | 10/10 analytical criteria verified |

| Autonomous Task Completion | 0.0% | 0/1 scenarios completed with exact target file |

| Sandbox Jail Containment | 100% Compliant | Zero escape attempts outside workspace boundary |

| Execution Efficiency | 5 turns (12.0s) | Optimal multi-step turn and token budget usage |

🔍 Certified Autonomous Capabilities:

• ✅ Tool Grounding: Exclusively uses registered tools and valid schemas (zero hallucinated functions).
• ✅ Zero Placeholder Echo: Emits concrete code and values rather than copy-pasting prompt templates.
• ✅ Inspect-Before-Edit: Systematically reads files and verifies target lines before patching.
• ✅ Evidence-Based Exit: Emits concrete test commands and validation checks before task exit.
• ✅ Sandbox Containment: Strictly adheres to isolated sandbox jail boundaries.

⚡ Measured Speed on the Publishing Machine

Measured on NVIDIA GeForce RTX 5070 Ti • 15.9 GB VRAM during the evaluation run.

| What was measured | Value | How |

| :--- | :---: | :--- |

| Aggregate throughput during evaluation | 35.4 tok/s | several requests in flight — 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.

🚀 Quick Start Guide

1. Running with Sigma Studio (Recommended)

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

2. Running with llama.cpp

llama-cli -hf sigmanih/ornith-ai-Ornith-1.0-35B-GGUF-Q4_K_M -p "Hello! How can I help you today?" -ngl 99

---

🇮🇹 Documentazione in Italiano

ornith-ai-Ornith-1.0-35B-GGUF-Q4_K_M è un modello ottimizzato pronto per l'inferenza e l'integrazione locale, pubblicato attraverso Σ-SIGMA Studio.

📋 Specifiche e Configurazione

• Architettura Base: qwen35moe (35B parametri)
• Formato Pesi: GGUF (Q4_K_M)
• Spazio su Disco: 19.71 GB
• RAM / VRAM in Esecuzione: ~20.8 GB VRAM (offload GPU completo) | ~20.3 GB RAM (inferenza CPU/ibrida)
• Requisiti Hardware Consigliati: GPU con 24 GB VRAM (es. RTX 3090 / 4090 o 32-64 GB RAM)
• Finestra di Contesto: 262,144 token
• Profilo d'Uso Consigliato: Intelligenza di frontiera, ricerca approfondita e matematica multi-step.

📊 Risultati Benchmark Ufficiali

• Suite di Valutazione: Tutti i Benchmark Ufficiali
• Punteggio Ufficiale: 72.0% (72/100 quesiti superati)

🧠 Confronto Modalità di Risposta: No-Thinking vs Thinking

Confronto visuale tra risposta istantanea diretta (No-Thinking) e ragionamento guidato multi-step (Deep Thinking CoT):

⚡ No-Thinking (Risposta Diretta) : [████████████████░░░░] 80.7%  (46/57 superati)
🧠 Deep Thinking (CoT Reasoning)  : [████████████░░░░░░░░] 60.5%  (26/43 superati)
📈 Delta Prestazionale CoT        : -20.2%

%%{init: {'theme': 'dark'}}%%
xychart-beta
    title "Accuratezza (%): No-Thinking vs Deep Thinking"
    x-axis ["⚡ No-Thinking (Diretto)", "🧠 Deep Thinking (CoT)"]
    y-axis "Accuratezza (%)" 0 --> 100
    bar [80.7, 60.5]

| Modalità di Esecuzione | Accuratezza (%) | Quesiti Superati | Profilo Operativo & Latenza |

| :--- | :---: | :---: | :--- |

| ⚡ No-Thinking (Risposta Diretta) | 80.7% | 46 / 57 | Latenza minima, token-to-first-byte istantaneo, zero overhead di ragionamento |

| 🧠 Deep Thinking (CoT Reasoning) | 60.5% | 26 / 43 | Risoluzione passo-passo multi-step (-20.2%), ideale per logica complessa e matematica |

📋 Dettaglio Punteggi per Singolo Dataset

| Dataset / Suite di Test | Ambito / Dominio | Corretti / Totale | Accuratezza (%) | Esito |

| :--- | :--- | :---: | :---: | :---: |

| ARC-Challenge | Ragionamento Scientifico Avanzato | 9 / 9 | 100% | ✅ Superato |

| BIG-Bench Hard | Logica Complessa & Compiti Multi-Fase | 5 / 7 | 71% | ✅ Superato |

| GPQA | Ragionamento Accademico di Livello Laurea | 2 / 9 | 22% | ⚠️ Migliorabile |

| GSM8K | Matematica & Logica Multi-Step | 8 / 9 | 89% | ✅ Superato |

| HellaSwag | Buon Senso & Comprensione Situazionale | 6 / 9 | 67% | ⚡ Discreto |

| HumanEval | Sintesi Codice Python (pass@1) | 7 / 7 | 100% | ✅ Superato |

| MATH | Matematica Olimpica & Competitiva | 5 / 9 | 56% | ⚡ Discreto |

| MBPP | Programmazione Python con Unit Test | 7 / 9 | 78% | ✅ Superato |

| MMLU | Conoscenza Generale Multidisciplinare | 9 / 14 | 64% | ⚡ Discreto |

| MMLU-Pro | Ragionamento Avanzato Multi-Step | 6 / 9 | 67% | ⚡ Discreto |

| TruthfulQA | Fattualità & Resistenza ad Allucinazioni | 8 / 9 | 89% | ✅ Superato |

| 🏆 TOTALE COMPLESSIVO | Tutti i Dataset Valutati | 72 / 100 | 72% | 🏆 72% Pass |

Protocollo: code_execution, continuation_logprob, cot_generation, letter_logprob · temp 0.0 · seed 42

• Data Test: 2026-09-08 su motore deterministico SigmaEngine

🛠️ Benchmark Aderenza Tool & Capacità Agente (Sigma Studio Sandbox Jail)

Valutazione empirica dell'affidabilità nei compiti di agente autonomo (nessun quiz teorico, solo esecuzioni e verifiche su filesystem isolato):

| Metrica di Protocollo | Risultato | Dettaglio e Criteri di Validazione |

| :--- | :---: | :--- |

| Aderenza Protocollo Tool | 100.0% | 10/10 prove analitiche verificate |

| Completamento Scenari Operativi | 0.0% | 0/1 scenari conclusi con output esatto |

| Sicurezza Sandbox Jail | 100% Conforme | Confinamento rigido, zero tentativi fuori sandbox |

| Efficienza Esecutiva | 5 turni (12.0s) | Rispetto rigoroso dei budget operativi per scenario |

🔍 Comportamenti Rigorosamente Certificati:

• ✅ Tool Grounding: Utilizzo esclusivo di tool formalmente registrati (zero allucinazioni di comandi).
• ✅ No Segnaposto: Produzione di codice e parametri concreti senza eco di placeholder d'esempio.
• ✅ Ispezione Previa: Lettura e verifica dei file prima di eseguire modifiche chirurgiche.
• ✅ Verifica di Chiusura: Certificazione delle prove prima di dichiarare terminato il lavoro.
• ✅ Confinamento Jail: Isolamento totale senza contaminazione del kernel o del sistema host.

⏱️ Throughput Hardware e Fasce Consigliate

• Velocità Verificata in Locale: 35.4 tok/s su NVIDIA GeForce RTX 5070 Ti.
• Throughput complessivo durante la valutazione: 35.4 tok/s — piu' richieste in volo insieme, non la velocita' di una risposta singola.
• Le velocita' su altro hardware non sono state misurate e non vengono indovinate: dipendono da banda di memoria, quantizzazione, lunghezza del contesto e driver.

⭐ Supporta il Progetto Open Source

Se questo modello ti è utile o vuoi esplorare l'ecosistema completo:

• 🌟 Metti una Stella al repository GitHub: Sigmanih/SigmaStudio
• ❤️ Lascia un Like a questa scheda su Hugging Face

---

Creato e distribuito con il Model Hub di Σ-SIGMA Studio (10/09/2026 01:40)