Downloads · 30 days
12
100% of all-time downloads
LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3
DuoNeural/LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3
LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3 is a text generation model from DuoNeural. Use it when you need the model to write or continue text. It is set up for hermes. The card lists the license as other.
DuoNeural/LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3 is the latest, apex-tier autonomous agentic coding model trained by DuoNeural (Aura ✨, Archon, and Jesse).
All-time downloads
12
Public
Parameters
8.5B
17 GB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors16.9 GB · 100%
Parameter types
How the weights are stored.
BF168.5B · 100%
Try a prompt
Base models
Model card
From the Hugging Face model README
DuoNeural-HYPERLFM-2.5-8B-Hermes-Agentic-Coder-Abliterated-v3 ✨
DuoNeural/LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3 is the latest, apex-tier autonomous agentic coding model trained by DuoNeural (Aura ✨, Archon, and Jesse).
Built upon our abliterated hybrid state-space & mixture-of-experts foundation architecture (DuoNeural/LFM2.5-8B-A1B-Abliterated), v3 elevates autonomous code generation, multi-step agentic planning, and contract-hardened logic by activating only 1.5 billion parameters per token out of its 8.3 billion total parameters.
It delivers blistering inference speeds (~360 tokens/sec on modern desktop/workstation GPUs and ~80–90 tokens/sec on legacy mobile GPUs like the GTX 1070) while breaking the 50.0% Pass@1 barrier on the rigorous EvalPlus HumanEval+ benchmark.
- 📦 GGUF Quantizations: DuoNeural/LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3-GGUF
- 🎯 PEFT LoRA Adapter: DuoNeural/LFM2.5-8B-A1B-Hermes-Agentic-Coder-Abliterated-v3-LoRA
- 🏛️ Previous Releases: v2 Model | v1 Model
🚀 Key Advancements in v3
-
Expanded Context Horizon (3,072 Tokens):
- +50% sequence length expansion over v2 (2,048 tokens), enabling deep multi-turn agentic loops, long file inspections, and comprehensive function declarations without truncation.
-
Expanded 65,185-Sample Curated Dataset Diet:
- Agentic Function Calling: 13,185 verified tool execution schemas (
NousResearch/hermes-function-calling-v1+Jofthomas/hermes-function-calling-thinking-V1). - Step-by-Step Reasoning & Chain of Thought: 18,000 deep reasoning traces (
bespokelabs/Bespoke-Stratos-17k+AI-MO/NuminaMath-CoT). - Contract-Hardened Code Generation & Debugging: 18,000 algorithmic edge-case & compiler-feedback repair pairs (
theblackcat102/evol-codealpaca-v1+m-a-p/CodeFeedback-Filtered-Instruction). - Fluid Multi-Turn Instruction: 16,000 natural conversation and instruction samples (
HuggingFaceTB/smoltalk).
- Agentic Function Calling: 13,185 verified tool execution schemas (
-
Dual-Mode Inference Dynamics:
- Direct Fast Mode: Conversational queries and straightforward code requests respond immediately at maximum throughput without bloated internal monologues.
- System 2
<thought>Mode: Mathematical reasoning, complex algorithmic problem solving, and explicit thinking prompts dynamically engage deep<thought>...</thought>chains of thought before concluding.
-
100% Uncensored Abliteration:
- Completely abliterated alignment layers permit unrestricted research into low-level kernel C, memory management, binary analysis, and penetration testing scenarios.
📈 Benchmark Evolution: Stock LFM vs. v2 vs. v3
Evaluated using standard zero-shot greedy decoding on the raw OpenAI-compatible API endpoint:
| Benchmark / Capability | Stock LFM 2.5 8B A1B | DuoNeural v2 QLoRA | DuoNeural v3 QLoRA | Delta (v3 vs Stock) |
|---|---|---|---|---|
| EvalPlus: HumanEval (Base) | ~36.8% | 52.4% | 56.1% Pass@1 (92/164) | +19.3% 🚀 |
| EvalPlus: HumanEval+ (Extra Tests) | ~31.2% | 46.3% | 50.0% Pass@1 (82/164) | +18.8% 🛡️ |
| EvalPlus: MBPP (Base) | ~45.0% | 59.3% | 60.8% Pass@1 (230/378) | +15.8% 📈 |
| EvalPlus: MBPP+ (Extra Tests) | ~38.1% | 48.9% | 49.7% Pass@1 (188/378) | +11.6% 🛡️ |
| Hermes Function Calling AST | 49.7% | 100.0% | 100.0% (25/25) | +50.3% 🛠️ |
| EOS Anomaly / Freeze Rate | ~50–70% drop | 0.0% | 0.0% (0/3) | Flawless 🎯 |
| Abliteration / Uncensored | Liquid AI Guardrails | 100.0% | 100.0% Uncensored | Uncensored 🔓 |
| Context Window | 32,768 | 2,048 | 3,072 tokens | +50% over v2 |
| Active Parameter Cost | 1.5B active | 1.5B active | 1.5B active | Maximum Efficiency |
🏆 Industry Benchmark Comparison (8B Parameter Class)
Comparing DuoNeural v3 against top industry models in the 7B–8B parameter tier. Note that while competitor models activate 7.0B–8.0B parameters on every token, DuoNeural v3 activates only 1.5B parameters per token:
| Model | Active Params | Total Params | HumanEval (Base) | HumanEval+ (Hard) | MBPP (Base) | MBPP+ (Hard) |
|---|---|---|---|---|---|---|
| DuoNeural LFM-2.5-8B v3 | 1.5B | 8.3B | 56.1% | 50.0% | 60.8% | 49.7% |
| DuoNeural LFM-2.5-8B v2 | 1.5B | 8.3B | 52.4% | 46.3% | 59.3% | 48.9% |
| Granite-3.3-8B-Instruct | 8.0B | 8.0B | 53.7% | 48.2% | 61.9% | 50.8% |
| Llama-3-8B-Instruct | 8.0B | 8.0B | 51.2% | 45.7% | 61.1% | 50.0% |
| Gemma-7B-it | 7.0B | 7.0B | 50.6% | 44.5% | 54.5% | 44.4% |
| Mistral-7B-Instruct-v0.3 | 7.0B | 7.0B | 42.1% | 37.8% | 50.8% | 41.3% |
| Stock LFM 2.5 8B A1B | 1.5B | 8.3B | 36.8% | 31.2% | 45.0% | 38.1% |
💡 How to Use & Prompting Dynamics
1. Direct vs. Reasoning Mode
- Fast / Agentic Execution: For coding questions, tool execution, or conversation, ask normally. The model responds directly with zero unnecessary delay:
User: Write a Python function to compute the SHA-256 hash of a file. Assistant: [Direct, high-speed Python code] - System 2 Deep Reasoning Mode: For difficult math, algorithm puzzles, or step-by-step proofs, prompt the model to think in
<thought>tags:System: You are an expert reasoning assistant. Think carefully through the logic in <thought> tags before answering. User: Prove that the square root of 2 is irrational. Assistant: <thought> [Step-by-step mathematical derivation...] </thought> [Final structured response]
2. LM Studio / llama.cpp Setup
- LM Studio: Load the
Q4_K_MGGUF. Set Context Length to3072(or up to16384with YaRN). If you wish to enable collapsible thinking, ensure the thinking prompt regex matches<thought>([\s\S]*?)</thought>. - Inference Speed: Expect ~350+ tps on RTX 4080/3090 workstations and ~80–90 tps on mobile laptops (GTX 1070).
👥 Authorship & Research Lab
Developed and released by DuoNeural Research Lab:
- Aura ✨ — Apex AI Cognitive Architect & Systems Co-Director
- Archon — Lead Theoretical Scientist & Quantum/Thermodynamics Architect
- Jesse — Founder, Visionary, and Systems Architect
DuoNeural is dedicated to open-source neuro-symbiotic machine intelligence, topological state-space computing, and sovereign uncensored AI.
- Task
- Text Generation
- Library
- hermes
- Type
- lfm2_moe
- License
- other
- Created
- Sep 18, 2026
- Updated
- Sep 19, 2026