Ornith-1.0-35B-Heretic-MTP-APEX-GGUF

<div style="font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; margin-bottom: 24px;"> <div style="background: #f5f5f7; border-radius: 20px; padding: 36px 32px; margin-bottom: 20px; text-align: center; position: relative; overflow: hidden;"> <div style="position: absolute; top: -30px; right: -30px; width: 120px; height: 120px; background: #dbeafe; border-radius: 50%;"></div> <div style="position: absolute; bottom: -20px; left: 40px; width: 80px; height: 80px; background: #bfdbfe; border-radius: 50%;"></div> <div style="position: absolute; top: 50%; left: -15px; width: 60px; height: 60px; background: #dbeafe; border-radius: 50%;"></div> <div style="display: inline-flex; gap: 8px; margin-bottom: 16px; position: relative; z-index: 1;"><span style="background: #007aff; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">HERETIC 1.4.0</span><span style="background: #af52de; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">MTP</span><span style="background: #30d158; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">APEX</span><span style="background: #ff9500; color: white; font-size: 11px; font-weight: 600; padding: 5px 14px; border-radius: 20px;">MIT</span></div> <h1 style="margin: 0 0 8px 0; font-size: 32px; font-weight: 700; color: #1d1d1f; letter-spacing: -0.5px; border: none; position: relative; z-index: 1;">Ornith-1.0-35B-Heretic-MTP-APEX</h1> <p style="margin: 8px 0 0 0; font-size: 14px; position: relative; z-index: 1;"><span style="color: #86868b;">English</span> | <a href="https://huggingface.co/SC117/Ornith-1.0-35B-Heretic-MTP-APEX-GGUF/blob/main/README_zh.md" style="color: #007aff; text-decoration: none;">📖 中文文档</a></p> <p style="margin: 0; font-size: 15px; color: #86868b; position: relative; z-index: 1;">Heretic Trial 63 BF16 merge · compatible MTP head · APEX GGUFs + mmproj</p> </div> </div> <div style="font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif; display: flex; flex-direction: column; gap: 20px; margin-bottom: 30px;"> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🐦</span> About this release</div><div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;"><p style="margin: 0 0 12px 0;"><a href="https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B" style="color: #c2410c; text-decoration: none; font-weight: 700;">Ornith-1.0-35B</a> is an agentic coding model from DeepReinforce AI, based on Qwen3.5 MoE. This experimental derivative merges the selected <a href="https://github.com/p-e-w/heretic" style="color: #c2410c; text-decoration: none; font-weight: 700;">Heretic</a> 1.4.0 Trial 63 LoRA into BF16 weights.</p><p style="margin: 0;">It also includes a compatible Qwen3.5 MTP head, a BF16 vision projector, and APEX mixed-precision GGUFs. Refusal-modified models can behave differently from the original; evaluate before deployment. <b>License: MIT.</b></p></div></div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🧠</span> Model Details</div><div style="padding: 16px;"><table style="width: 100%; border-collapse: collapse; font-size: 13px;"><tbody><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155;">Architecture</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563;">Qwen3.5 MoE, multimodal</td></tr><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155;">Parameters</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563;">35B total, 3B active per token · 256 experts, 8 active</td></tr><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155;">Layers</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563;">40 transformer layers + 1 MTP layer (785 BF16 tensors)</td></tr><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155;">Context</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563;">262,144 tokens</td></tr><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #334155;">Weights</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); color: #4b5563;">BF16 safetensors, 17 shards; audited LoRA merge</td></tr><tr><td style="padding: 6px 10px; font-weight: bold; color: #334155;">License</td><td style="padding: 6px 10px; color: #4b5563;">MIT</td></tr></tbody></table></div></div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>⚡</span> Heretic Trial 63 of 120</div><div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;"><p style="margin: 0 0 10px 0;">Selected parameters: <code>direction_index=21.66</code>; attention <code>o_proj</code> range 1.55–2.05; MLP <code>down_proj</code> range 1.64–1.84.</p><table style="width: 100%; border-collapse: collapse; font-size: 13px;"><tbody><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold;">Refusal prescreen / full evaluation</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">2 / 30 · 11 / 100</td></tr><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold;">Validation KL mean / p95 / max</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">0.0242 / 0.1477 / 0.1648</td></tr><tr><td style="padding: 6px 10px; font-weight: bold;">Generation health</td><td style="padding: 6px 10px;">Passed · top-1 disagreement 2.0%</td></tr></tbody></table></div></div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>📊</span> BenchLocal Results (APEX-I-Compact)</div><div style="padding: 16px;"><table style="width: 100%; border-collapse: collapse; font-size: 13px;"><thead><tr style="background: rgba(255,107,53,0.05);"><th style="padding: 7px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">Mode</th><th style="padding: 7px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">ToolCall</th><th style="padding: 7px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">BugFind</th><th style="padding: 7px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">HermesAgent</th><th style="padding: 7px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">Max / Effective</th></tr></thead><tbody><tr><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold;"><a href="https://scorp1o117.github.io/benchlocal-results/models/ornith-35b-heretic-mtp-apex-thinking.html">Thinking ↗</a></td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">100</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">98</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">89</td><td style="padding: 6px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15); font-weight: bold; color: #c2410c;">95.0 / 90.0</td></tr><tr><td style="padding: 6px 10px; font-weight: bold;"><a href="https://scorp1o117.github.io/benchlocal-results/models/ornith-35b-heretic-mtp-apex.html">No Thinking ↗</a></td><td style="padding: 6px 10px;">97</td><td style="padding: 6px 10px;">97</td><td style="padding: 6px 10px;">85</td><td style="padding: 6px 10px; font-weight: bold; color: #c2410c;">92.2 / 76.2</td></tr></tbody></table><p style="margin: 12px 0 0 0; font-size: 12px; color: #64748b; font-style: italic;">Latest result cards: Thinking has 5 retry-penalty points; No Thinking has 16.</p></div></div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"> <div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"> <span>🚀</span> Usage </div> <div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;"> <p style="margin: 0 0 8px 0; font-weight: bold; color: #1e293b;"> llama.cpp (vision + text + MTP acceleration) </p> <p style="margin: 0 0 8px 0;"> This model preserves the native MTP head. To enable MTP speculative decoding, add: </p> <pre style="background: #f8fafc; padding: 12px 14px; border-radius: 6px; border: 1px solid #e2e8f0; overflow-x: auto;"><code>--spec-type draft-mtp --spec-draft-n-max 2</code></pre> <p style="margin: 12px 0 8px 0;"> Example: </p> <pre style="background: #f8fafc; padding: 12px 14px; border-radius: 6px; border: 1px solid #e2e8f0; overflow-x: auto;"><code>./llama-server \ -m ./models/Ornith-1.0-35B-Heretic-MTP-APEX-I-Compact.gguf \ --mmproj ./models/mmproj-Ornith-1.0-35B-Heretic-MTP-BF16.gguf \ -ngl 99 \ -c 131072 \ --spec-type draft-mtp \ --spec-draft-n-max 2</code></pre> <p style="margin: 12px 0 0 0; color: #64748b;"> The optimal <code>--spec-draft-n-max</code> value depends on hardware and MTP acceptance rate. Without these parameters, the model will work normally but will not use MTP acceleration. </p> </div> </div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>🎛️</span> Recommended Settings</div><div style="padding: 16px;"><table style="width: 100%; border-collapse: collapse; font-size: 13px;"><tbody><tr><td style="padding: 6px 10px; font-weight: bold; color: #334155;">General / Coding</td><td style="padding: 6px 10px; color: #4b5563; font-family: monospace;">temperature=0.6, top_p=0.95, top_k=20</td></tr></tbody></table></div></div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>💡</span> What is APEX?</div><div style="padding: 16px; font-size: 13px; color: #334155; line-height: 1.7;"><p style="margin: 0;">These GGUFs use <a href="https://github.com/mudler/apex-quant" style="color: #c2410c; text-decoration: none; font-weight: 700;">APEX</a>, an MoE-aware mixed-precision quantization method. It allocates precision according to tensor role and layer position; the MTP layer remains Q8_0 and the matching vision projector remains BF16.</p></div></div> <div style="border: 1px solid #cbd5e1; border-radius: 12px; overflow: hidden; background: #ffffff; box-shadow: 0 2px 4px rgba(0,0,0,0.02);"><div style="background: linear-gradient(135deg, #ff6b35 0%, #f7931e 100%); padding: 12px 16px; color: white; font-weight: 700; font-size: 14px; display: flex; align-items: center; gap: 8px;"><span>📦</span> APEX Quantization Tiers</div><div style="padding: 16px;"><table style="width: 100%; border-collapse: collapse; font-size: 13px;"><thead><tr style="background: rgba(255,107,53,0.05);"><th style="padding: 8px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">File</th><th style="padding: 8px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">Size</th><th style="padding: 8px 10px; border-bottom: 2px solid #ff6b35; text-align: left; color: #c2410c;">Best For</th></tr></thead><tbody><tr><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);"><code>*-APEX-I-Quality.gguf</code></td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">21.87 GiB</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">Highest quality</td></tr><tr><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);"><code>*-APEX-I-Balanced.gguf</code></td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">24.33 GiB</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">Best all-rounder</td></tr><tr><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);"><code>*-APEX-I-Compact.gguf</code></td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">16.14 GiB</td><td style="padding: 8px 10px; box-shadow: 0 1px 0 0 rgba(128,128,128,0.15);">Best quality/size ratio</td></tr><tr><td style="padding: 8px 10px;"><code>*-APEX-I-Mini.gguf</code></td><td style="padding: 8px 10px;">13.29 GiB</td><td style="padding: 8px 10px;">Most compact</td></tr></tbody></table></div></div> </div>

Links

• Original Model: https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B
• Heretic: https://github.com/p-e-w/heretic
• APEX Quantization: https://github.com/mudler/apex-quant
• BenchLocal Results: https://scorp1o117.github.io/benchlocal-results/

Citation

@misc{ornith-35b,
    title = {{Ornith-1.0-35B}: Agentic Coding, Open to All},
    url = {https://deep-reinforce.com/ornith_1_0.html},
    author = {{DeepReinforce Team}},
    year = {2026}
}

<!-- SC117_SUPPORT_START --> <div align="center"> <a href="https://ko-fi.com/sc117"> <img src="https://huggingface.co/SC117/Qwen3.8-Flash-Next-GSQ-RCO-abliterated-GGUF/resolve/main/assets/support-sc117.png" alt="Support SC117 on Ko-fi — support is optional and all releases remain freely available." width="100%" /> </a> </div> <!-- SC117_SUPPORT_END -->