{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b","slug":"nousresearch-hermes-3-llama-3.1-70b","name":"Nous: Hermes 3 70B Instruct","type":"model","url":"https://openrouter.ai/models/nousresearch~hermes-3-llama-3.1-70b","page_url":"https://unfragile.ai/nousresearch-hermes-3-llama-3.1-70b","categories":["chatbots-assistants"],"tags":["nousresearch","api-access","text"],"pricing":{"model":"paid","free":false,"starting_price":"$3.00e-7 per prompt token"},"status":"active","verified":false},"capabilities":[{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_0","uri":"capability://text.generation.language.multi.turn.conversational.reasoning.with.extended.context.coherence","name":"multi-turn conversational reasoning with extended context coherence","description":"Hermes 3 70B maintains semantic coherence across extended multi-turn conversations through optimized attention mechanisms and training on long-context datasets, enabling it to track conversation state, reference earlier turns accurately, and resolve pronouns/references across 10+ exchanges without context collapse. The model uses Llama 3.1's grouped-query attention (GQA) architecture to reduce KV cache memory while preserving long-range dependencies, allowing it to handle conversations that would cause context drift in smaller models.","intents":["Build a multi-turn chatbot that remembers context across 20+ exchanges without losing coherence","Create conversational agents that can reference earlier discussion points accurately","Develop customer support systems that maintain conversation history without degradation"],"best_for":["Teams building stateful conversational AI systems","Developers creating long-form dialogue applications","Enterprises needing reliable multi-turn customer interactions"],"limitations":["Context window is finite (likely 8K-128K tokens depending on deployment); very long conversations still require external memory/summarization","Attention mechanisms add computational overhead; inference latency increases with conversation length","No built-in conversation state persistence — requires external session management"],"requires":["API access to OpenRouter or compatible inference endpoint","Conversation history management system (in-memory or database)","Token counting library to track context usage"],"input_types":["text (natural language user messages)","structured conversation history (turn-by-turn format)"],"output_types":["text (natural language responses)","structured dialogue acts (if prompted for structured output)"],"categories":["text-generation-language","conversational-ai"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_1","uri":"capability://tool.use.integration.agentic.tool.use.orchestration.with.function.calling","name":"agentic tool-use orchestration with function calling","description":"Hermes 3 70B is trained to generate structured function calls in response to tool-use prompts, enabling it to invoke external APIs, execute code, or trigger workflows by outputting properly-formatted JSON or XML function signatures. The model learns to reason about which tools to invoke, in what order, and with what parameters through instruction-tuning on synthetic agentic datasets, allowing it to decompose complex tasks into tool-calling sequences without requiring explicit prompt engineering for each tool.","intents":["Build autonomous agents that can call APIs, databases, or custom functions to complete multi-step tasks","Create code-execution agents that can write and invoke Python/JavaScript snippets","Develop workflow automation systems where the model decides which tools to chain together"],"best_for":["Developers building LLM agents with external tool dependencies","Teams creating autonomous workflow systems","Builders prototyping multi-step task automation"],"limitations":["Tool-calling accuracy degrades with >10 available tools; model may hallucinate function names or parameters","Requires careful prompt engineering to define tool schemas; ambiguous schemas lead to incorrect function calls","No native error handling or retry logic — agent framework must implement fallback strategies"],"requires":["Tool/function schema definitions (JSON or XML format)","Agent framework to parse function calls and execute them (e.g., LangChain, LlamaIndex, custom)","External tools/APIs to invoke"],"input_types":["text (natural language task description)","structured tool schemas (JSON/XML function definitions)"],"output_types":["structured function calls (JSON/XML)","text (reasoning about which tools to use)"],"categories":["tool-use-integration","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_10","uri":"capability://search.retrieval.semantic.search.and.relevance.ranking.over.custom.knowledge.bases","name":"semantic search and relevance ranking over custom knowledge bases","description":"Hermes 3 70B can be used as a semantic understanding layer to rank the relevance of documents or passages to a query by understanding semantic similarity and contextual relevance, enabling it to identify the most relevant information from a knowledge base without requiring explicit vector embeddings. The model learns to understand query intent and match it against document content based on meaning rather than keyword matching, enabling more intelligent search and retrieval.","intents":["Build semantic search systems that understand query intent beyond keywords","Create document ranking systems that identify most relevant passages","Develop knowledge base systems that retrieve contextually relevant information"],"best_for":["Teams building semantic search systems","Developers creating knowledge base retrieval systems","Enterprises needing intelligent document ranking"],"limitations":["Ranking quality depends on document quality and relevance; garbage in, garbage out","Inference latency increases with knowledge base size; not suitable for real-time search over millions of documents without external indexing","No built-in vector embeddings; requires external embedding model or re-ranking approach"],"requires":["Query text","Knowledge base or document collection to search","Optional: external vector embedding system for pre-filtering"],"input_types":["text (search query)","text (documents to rank)"],"output_types":["text (ranked documents or passages)","structured data (relevance scores if formatted)"],"categories":["search-retrieval","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_2","uri":"capability://text.generation.language.advanced.roleplay.and.character.consistency","name":"advanced roleplay and character consistency","description":"Hermes 3 70B maintains consistent character personas, voice, and behavioral patterns across extended interactions through instruction-tuning on roleplay datasets and character-consistency examples. The model learns to internalize character traits, speech patterns, and knowledge domains, allowing it to stay in-character while responding contextually to user inputs without breaking character or contradicting established persona attributes.","intents":["Build interactive fiction or game systems with consistent NPC characters","Create educational tutoring systems where the tutor maintains a consistent teaching persona","Develop entertainment chatbots with distinct personalities that don't drift or contradict themselves"],"best_for":["Game developers building NPC dialogue systems","Educational content creators building character-based tutors","Entertainment platforms requiring consistent character interactions"],"limitations":["Character consistency degrades over very long conversations (100+ turns); periodic character re-prompting needed","Complex multi-character scenarios may cause character bleed (one character adopting traits of another)","Requires explicit character definition in system prompt; vague character descriptions lead to inconsistent behavior"],"requires":["Detailed character definition (background, personality, speech patterns, knowledge domain)","System prompt engineering to establish character context","Conversation history to maintain character state"],"input_types":["text (user dialogue input)","structured character definitions (JSON or prose)"],"output_types":["text (character-consistent dialogue responses)"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_3","uri":"capability://planning.reasoning.structured.reasoning.and.chain.of.thought.decomposition","name":"structured reasoning and chain-of-thought decomposition","description":"Hermes 3 70B is trained to generate explicit reasoning chains where it breaks down complex problems into intermediate steps, showing its work before arriving at conclusions. The model learns to use natural language reasoning tokens (e.g., 'Let me think through this step by step...') and structured formats to decompose problems, enabling more reliable multi-step reasoning and making its decision-making process interpretable to users and downstream systems.","intents":["Build systems that need to show reasoning steps for explainability or debugging","Create math/logic problem solvers that break down solutions into verifiable steps","Develop decision-support systems where intermediate reasoning is auditable"],"best_for":["Teams building explainable AI systems","Developers creating educational problem-solving assistants","Enterprises requiring auditable AI decision-making"],"limitations":["Chain-of-thought reasoning increases token output by 2-5x, raising inference costs and latency","Reasoning chains can contain logical errors that aren't caught by the model; external verification needed","Model may generate plausible-sounding but incorrect intermediate steps (reasoning hallucination)"],"requires":["Prompts that explicitly request step-by-step reasoning","Token budget for longer outputs (reasoning adds overhead)","Optional: verification system to validate intermediate steps"],"input_types":["text (problem statement or question)"],"output_types":["text (reasoning chain + final answer)","structured reasoning steps (if formatted with delimiters)"],"categories":["planning-reasoning","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_4","uri":"capability://code.generation.editing.code.generation.and.completion.with.multi.language.support","name":"code generation and completion with multi-language support","description":"Hermes 3 70B generates syntactically correct code across 40+ programming languages (Python, JavaScript, Java, C++, Go, Rust, etc.) through training on diverse code repositories and instruction-tuning on code-generation tasks. The model understands language-specific idioms, libraries, and best practices, allowing it to generate production-ready code snippets, complete partial implementations, and suggest refactorings with language-aware context awareness.","intents":["Generate boilerplate code or function implementations from natural language descriptions","Complete partial code snippets with context-aware suggestions","Translate code between programming languages while preserving logic"],"best_for":["Developers using code generation to accelerate development","Teams building code-generation tools or IDE plugins","Educators creating coding tutorials with AI assistance"],"limitations":["Generated code may contain logical errors or security vulnerabilities; requires human review before production use","Performance degrades on domain-specific languages or less common languages (e.g., Cobol, Lisp)","No built-in linting or syntax validation; output requires testing"],"requires":["Language specification in prompt (e.g., 'Python 3.9')","Optional: code context or existing codebase for better completion","Testing/validation framework to verify generated code"],"input_types":["text (natural language code description or partial code)","code (existing code context for completion)"],"output_types":["code (generated or completed code snippets)","text (explanations of generated code)"],"categories":["code-generation-editing","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_5","uri":"capability://planning.reasoning.instruction.following.with.complex.task.decomposition","name":"instruction-following with complex task decomposition","description":"Hermes 3 70B is trained to follow detailed, multi-part instructions with high fidelity, parsing complex task specifications and executing them accurately even when instructions contain multiple constraints, conditional logic, or nested requirements. The model learns to clarify ambiguous instructions, ask for missing information, and decompose complex tasks into sub-steps, enabling it to handle real-world task specifications that aren't perfectly formatted.","intents":["Build systems that execute complex user instructions with multiple constraints","Create task automation systems where instructions are specified in natural language","Develop assistants that can handle ambiguous or incomplete task specifications"],"best_for":["Teams building instruction-following agents","Developers creating task automation platforms","Enterprises automating complex business processes"],"limitations":["Instruction-following accuracy degrades with >5 nested constraints or conditional branches","Model may misinterpret ambiguous instructions; requires clarification mechanisms","No built-in validation that instructions were followed correctly; requires external verification"],"requires":["Clear task specification (natural language or structured format)","Optional: examples of correct task execution for few-shot learning","Verification system to validate task completion"],"input_types":["text (natural language task instructions)"],"output_types":["text (task execution results or clarification questions)","structured data (if task output is structured)"],"categories":["planning-reasoning","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_6","uri":"capability://text.generation.language.knowledge.synthesis.and.summarization.with.context.preservation","name":"knowledge synthesis and summarization with context preservation","description":"Hermes 3 70B synthesizes information from multiple sources or long documents into coherent summaries while preserving key context, nuance, and important details. The model learns to identify salient information, abstract away redundancy, and maintain semantic relationships between concepts, enabling it to create summaries at various granularities (bullet points, paragraphs, abstracts) without losing critical information.","intents":["Summarize long documents or research papers into concise overviews","Extract key insights from multiple sources and synthesize them into coherent narratives","Create executive summaries of technical documentation or meeting transcripts"],"best_for":["Knowledge workers processing large volumes of text","Teams building document analysis systems","Researchers synthesizing literature reviews"],"limitations":["Summarization quality degrades on highly technical or domain-specific content without domain context","Model may omit important details if they're not explicitly highlighted in source material","Abstractive summarization can introduce subtle inaccuracies or misrepresentations"],"requires":["Source text or documents to summarize","Optional: summary length or format specification","Optional: domain context or key terms to preserve"],"input_types":["text (documents, articles, transcripts to summarize)"],"output_types":["text (summaries in various formats: bullet points, paragraphs, abstracts)"],"categories":["text-generation-language","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_7","uri":"capability://text.generation.language.creative.writing.and.content.generation.with.style.control","name":"creative writing and content generation with style control","description":"Hermes 3 70B generates original creative content (stories, poetry, marketing copy, dialogue) while maintaining consistent tone, style, and voice through instruction-tuning on diverse writing datasets. The model learns to adapt its writing style to match specified genres, audiences, or tones (formal, casual, humorous, etc.), enabling it to generate contextually appropriate content that aligns with user intent and brand voice.","intents":["Generate marketing copy or product descriptions with specific brand voice","Create story outlines or dialogue for interactive fiction or games","Write poetry or creative content in specified styles or genres"],"best_for":["Content creators and marketers","Game developers building narrative content","Agencies generating creative assets at scale"],"limitations":["Generated content may lack originality or contain clichés, especially for common genres","Style consistency degrades over very long outputs (1000+ words); periodic re-prompting needed","Creative quality is subjective; requires human editorial review"],"requires":["Style/tone specification (genre, audience, voice guidelines)","Optional: examples of desired writing style for few-shot learning","Human editorial review for quality assurance"],"input_types":["text (creative brief, outline, or style specification)"],"output_types":["text (creative content: stories, poetry, marketing copy, dialogue)"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_8","uri":"capability://text.generation.language.question.answering.with.source.attribution.and.uncertainty.quantification","name":"question-answering with source attribution and uncertainty quantification","description":"Hermes 3 70B answers questions based on provided context or its training knowledge while optionally attributing answers to specific sources and expressing uncertainty about answers it's less confident in. The model learns to distinguish between high-confidence factual answers and speculative responses, enabling it to provide nuanced answers that acknowledge knowledge gaps or ambiguity rather than hallucinating confident but incorrect answers.","intents":["Build QA systems that cite sources for answers","Create customer support systems that acknowledge when they don't know answers","Develop research assistants that distinguish between confident and uncertain responses"],"best_for":["Teams building QA systems requiring source attribution","Customer support platforms needing honest uncertainty handling","Research and knowledge-work applications"],"limitations":["Uncertainty quantification is implicit (via language cues) rather than explicit probabilities; requires interpretation","Model may still hallucinate confident-sounding but incorrect answers despite training for uncertainty","Source attribution requires context to be provided; works poorly with pure knowledge-based QA"],"requires":["Question input","Optional: context documents for source-based QA","Optional: knowledge base or retrieval system for knowledge-based QA"],"input_types":["text (questions)","text (optional context documents)"],"output_types":["text (answers with optional source citations and uncertainty expressions)"],"categories":["text-generation-language","search-retrieval"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-3-llama-3.1-70b__cap_9","uri":"capability://text.generation.language.translation.and.cross.lingual.understanding","name":"translation and cross-lingual understanding","description":"Hermes 3 70B translates text between 50+ languages while preserving meaning, tone, and cultural context through training on multilingual corpora and instruction-tuning on translation tasks. The model understands language-specific idioms, grammar structures, and cultural references, enabling it to produce natural translations rather than literal word-for-word conversions, and can also answer questions or perform tasks in non-English languages.","intents":["Translate documents or user-generated content between languages","Build multilingual chatbots that serve users in their native languages","Create cross-lingual search or content discovery systems"],"best_for":["Global teams needing translation services","Platforms serving multilingual user bases","Developers building international applications"],"limitations":["Translation quality varies significantly by language pair; high-resource pairs (English-Spanish) are better than low-resource pairs (English-Swahili)","Idioms and cultural references may not translate perfectly; requires human review for marketing/creative content","Model may struggle with technical terminology in non-English languages"],"requires":["Source language and target language specification","Text to translate","Optional: domain context or glossary for technical terms"],"input_types":["text (in any supported language)"],"output_types":["text (translated to target language)"],"categories":["text-generation-language","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":26,"verified":false,"data_access_risk":"high","permissions":["API access to OpenRouter or compatible inference endpoint","Conversation history management system (in-memory or database)","Token counting library to track context usage","Tool/function schema definitions (JSON or XML format)","Agent framework to parse function calls and execute them (e.g., LangChain, LlamaIndex, custom)","External tools/APIs to invoke","Query text","Knowledge base or document collection to search","Optional: external vector embedding system for pre-filtering","Detailed character definition (background, personality, speech patterns, knowledge domain)"],"failure_modes":["Context window is finite (likely 8K-128K tokens depending on deployment); very long conversations still require external memory/summarization","Attention mechanisms add computational overhead; inference latency increases with conversation length","No built-in conversation state persistence — requires external session management","Tool-calling accuracy degrades with >10 available tools; model may hallucinate function names or parameters","Requires careful prompt engineering to define tool schemas; ambiguous schemas lead to incorrect function calls","No native error handling or retry logic — agent framework must implement fallback strategies","Ranking quality depends on document quality and relevance; garbage in, garbage out","Inference latency increases with knowledge base size; not suitable for real-time search over millions of documents without external indexing","No built-in vector embeddings; requires external embedding model or re-ranking approach","Character consistency degrades over very long conversations (100+ turns); periodic character re-prompting needed","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.05,"quality":0.47,"ecosystem":0.24,"match_graph":0.25,"freshness":0.9,"weights":{"adoption":0.35,"quality":0.2,"ecosystem":0.1,"match_graph":0.3,"freshness":0.05}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-05-24T12:16:24.484Z","last_scraped_at":"2026-05-03T15:20:45.776Z","last_commit":null},"community":{"stars":null,"forks":null,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=nousresearch-hermes-3-llama-3.1-70b","compare_url":"https://unfragile.ai/compare?artifact=nousresearch-hermes-3-llama-3.1-70b"}},"signature":"287mzCN25hehjLwlqYH9tihpIKTNFWh/zN89ksJcZri5s5TNPCcYo5BowRRNq3UKUys68AC4Udo1Q+CV0GQ9AQ==","signedAt":"2026-06-17T00:13:19.131Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/nousresearch-hermes-3-llama-3.1-70b","artifact":"https://unfragile.ai/nousresearch-hermes-3-llama-3.1-70b","verify":"https://unfragile.ai/api/v1/verify?slug=nousresearch-hermes-3-llama-3.1-70b","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}