{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b","slug":"nousresearch-hermes-2-pro-llama-3-8b","name":"NousResearch: Hermes 2 Pro - Llama-3 8B","type":"model","url":"https://openrouter.ai/models/nousresearch~hermes-2-pro-llama-3-8b","page_url":"https://unfragile.ai/nousresearch-hermes-2-pro-llama-3-8b","categories":["model-training"],"tags":["nousresearch","api-access","text"],"pricing":{"model":"paid","free":false,"starting_price":"$1.40e-7 per prompt token"},"status":"active","verified":false},"capabilities":[{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_0","uri":"capability://text.generation.language.instruction.following.conversation.with.function.calling","name":"instruction-following conversation with function calling","description":"Hermes 2 Pro processes multi-turn conversations and generates contextually appropriate responses using a transformer-based architecture trained on the OpenHermes 2.5 dataset. The model supports structured function calling through JSON schema inference, allowing it to parse user intents and invoke external tools or APIs by generating properly formatted function calls within its response stream. Training on instruction-tuned data enables the model to follow complex, multi-step directives and maintain conversation coherence across extended contexts.","intents":["build a conversational AI agent that can call external APIs based on user requests","create a chatbot that understands nuanced instructions and follows them precisely","integrate a language model into an agentic workflow that requires tool invocation","deploy a chat interface that handles both dialogue and structured function execution"],"best_for":["developers building agentic systems with tool-calling requirements","teams deploying conversational AI with external API integration","builders prototyping multi-turn dialogue systems with function execution"],"limitations":["8B parameter size limits reasoning depth on highly complex multi-step problems compared to 70B+ models","function calling accuracy depends on clarity of schema definition; ambiguous schemas may produce malformed JSON","context window size (likely 8K tokens based on Llama-3 8B standard) constrains conversation history and document context","no guaranteed output format validation — requires downstream JSON parsing and error handling"],"requires":["OpenRouter API key or compatible OpenAI-format API endpoint","HTTP client library (curl, Python requests, JavaScript fetch, etc.)","JSON schema definition for function calling (OpenAI function_calling or similar format)","Handling for streaming or non-streaming response modes"],"input_types":["text (natural language instructions, questions, conversation turns)","structured prompts with function schemas (JSON)","multi-turn conversation history (array of messages)"],"output_types":["text (natural language responses)","JSON (function calls with arguments)","mixed streaming output (text + function calls interleaved)"],"categories":["text-generation-language","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_1","uri":"capability://code.generation.editing.codebase.aware.code.generation.and.completion","name":"codebase-aware code generation and completion","description":"Hermes 2 Pro generates code snippets, functions, and multi-file solutions by leveraging transformer attention over code context provided in the prompt. The model was trained on diverse code examples from the OpenHermes dataset, enabling it to understand programming language syntax, common patterns, and API conventions. Code generation works through next-token prediction with awareness of language-specific indentation, bracket matching, and semantic structure, allowing it to produce syntactically valid code across multiple languages.","intents":["generate boilerplate code or function implementations from natural language descriptions","complete partial code snippets with context-aware suggestions","translate code between programming languages or frameworks","generate test cases or documentation from existing code"],"best_for":["developers seeking a lightweight, locally-deployable code generation model","teams building IDE plugins or code editors with integrated AI assistance","builders prototyping code generation features without heavy infrastructure"],"limitations":["8B parameter size struggles with very long files (>2K lines) or complex architectural decisions requiring deep codebase understanding","no built-in awareness of project structure, dependencies, or type definitions — requires explicit context in prompt","generated code may contain logical errors or inefficiencies; always requires review and testing","no real-time incremental generation optimization — full response must be generated before streaming begins"],"requires":["OpenRouter API key or compatible endpoint","Code context provided in prompt (file snippets, function signatures, or full files)","Programming language specification in prompt for accurate syntax generation","Post-processing for syntax validation and linting"],"input_types":["text (natural language code requests)","code snippets (partial implementations to complete)","structured prompts with language and framework context"],"output_types":["code (single functions, classes, or multi-file solutions)","text (explanations of generated code)","mixed (code with inline comments or documentation)"],"categories":["code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_2","uri":"capability://text.generation.language.multi.language.translation.and.paraphrasing","name":"multi-language translation and paraphrasing","description":"Hermes 2 Pro translates text between natural languages and paraphrases content by leveraging transformer-based sequence-to-sequence capabilities trained on multilingual examples in the OpenHermes dataset. The model performs translation through attention mechanisms that map source language tokens to target language equivalents, maintaining semantic meaning and context. Paraphrasing works similarly, using the same language for both input and output while varying syntax and word choice to preserve intent.","intents":["translate user-generated content or documentation into multiple languages","paraphrase text for clarity, tone adjustment, or plagiarism avoidance","localize chatbot responses for international audiences","generate alternative phrasings of instructions or prompts for testing"],"best_for":["teams building multilingual applications or content platforms","developers creating international chatbots or support systems","content creators needing bulk paraphrasing or translation"],"limitations":["Translation quality varies by language pair; less common language combinations may produce lower-quality output","No specialized translation memory or terminology database — may produce inconsistent terminology across documents","Context window limitations (8K tokens) constrain document-level translation of long-form content","Paraphrasing may alter technical precision or domain-specific terminology unintentionally"],"requires":["OpenRouter API key","Source language specification in prompt (for translation)","Target language specification","Optional: terminology or style guide for consistency"],"input_types":["text (source language content)","structured prompts with language pair and tone specifications"],"output_types":["text (translated or paraphrased content)","structured data (translations with confidence scores if post-processed)"],"categories":["text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_3","uri":"capability://data.processing.analysis.structured.data.extraction.and.schema.based.output.generation","name":"structured data extraction and schema-based output generation","description":"Hermes 2 Pro extracts structured information from unstructured text and generates JSON or other structured formats by understanding schema definitions provided in prompts. The model uses instruction-tuning to follow format specifications, generating valid JSON objects that conform to specified schemas. Extraction works through attention over source text, identifying relevant information and mapping it to schema fields, with the model learning to handle missing data, type conversions, and nested structures through training examples.","intents":["extract entities (names, dates, amounts) from documents or user input into structured JSON","convert natural language descriptions into API request payloads","parse unstructured logs or text into tabular or hierarchical data structures","generate configuration files or data models from plain English specifications"],"best_for":["developers building data pipelines that ingest unstructured text","teams automating form filling or data entry from documents","builders creating ETL workflows with LLM-based extraction"],"limitations":["No schema validation — generated JSON may be malformed or not conform to specified schema; requires post-processing validation","Extraction accuracy depends on clarity of schema definition and source text quality","Complex nested schemas or large field counts may exceed model's ability to maintain consistency","No built-in error handling for ambiguous or missing data — model may hallucinate values or omit fields"],"requires":["OpenRouter API key","JSON schema or format specification in prompt","Source text or document content","JSON schema validation library for post-processing (jsonschema, Zod, etc.)"],"input_types":["text (unstructured source content)","structured prompts with JSON schema definitions","documents or logs in text format"],"output_types":["JSON (structured objects conforming to schema)","text (with embedded JSON)","CSV or other delimited formats (if specified in schema)"],"categories":["data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_4","uri":"capability://planning.reasoning.reasoning.and.step.by.step.problem.decomposition","name":"reasoning and step-by-step problem decomposition","description":"Hermes 2 Pro performs multi-step reasoning by generating intermediate reasoning steps (chain-of-thought) before producing final answers. The model was trained on examples that demonstrate step-by-step problem solving, enabling it to break down complex questions into smaller sub-problems, work through them sequentially, and synthesize results. This capability works through next-token prediction where the model learns to generate explicit reasoning tokens before final answers, improving accuracy on tasks requiring logical deduction, arithmetic, or multi-hop inference.","intents":["solve math problems or logic puzzles by showing work step-by-step","answer complex questions that require multi-hop reasoning or fact synthesis","debug code or troubleshoot issues by walking through problem systematically","explain decision-making processes in agentic workflows for transparency"],"best_for":["developers building reasoning-heavy agents or decision-making systems","teams creating educational or tutoring applications","builders needing explainable AI outputs with visible reasoning chains"],"limitations":["8B parameter size limits reasoning depth on very complex problems; struggles with >5-step reasoning chains","Reasoning quality degrades on out-of-distribution problems not well-represented in training data","No built-in verification of intermediate steps — generated reasoning may contain logical errors","Increased token generation (reasoning + answer) increases latency and API costs compared to direct answers"],"requires":["OpenRouter API key","Prompts explicitly requesting step-by-step reasoning (e.g., 'Think step by step')","Sufficient context window to accommodate reasoning output (8K tokens minimum)"],"input_types":["text (questions or problems requiring reasoning)","structured prompts with reasoning format specifications"],"output_types":["text (step-by-step reasoning followed by final answer)","structured reasoning chains (if post-processed into JSON)"],"categories":["planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_5","uri":"capability://text.generation.language.conversational.context.management.and.multi.turn.dialogue","name":"conversational context management and multi-turn dialogue","description":"Hermes 2 Pro maintains conversational state across multiple turns by processing message history as a sequence of alternating user and assistant messages. The model uses transformer attention to track context from previous exchanges, enabling it to reference earlier statements, maintain consistent persona, and build on prior responses. Context management works through prompt formatting where the entire conversation history is concatenated and fed to the model, with the model learning to attend to relevant prior messages while ignoring irrelevant ones through training on multi-turn dialogue examples.","intents":["build chatbots that remember context across multiple user messages","create conversational agents that reference earlier statements or decisions","maintain consistent tone and persona throughout extended conversations","implement dialogue systems that handle clarifications and follow-up questions"],"best_for":["developers building customer support chatbots or conversational interfaces","teams creating interactive AI assistants with persistent context","builders prototyping dialogue systems without complex state management"],"limitations":["Context window size (8K tokens) limits conversation history; older messages are lost when history exceeds window","No built-in conversation summarization — requires external summarization to preserve context beyond window limit","Attention mechanism may lose track of context in very long conversations (>50 turns) due to attention dilution","No persistent storage of conversation state — each API call requires full history to be sent"],"requires":["OpenRouter API key","Message history formatted as array of {role, content} objects","Client-side conversation state management (storing history between API calls)","Optional: conversation summarization for long-running sessions"],"input_types":["text (current user message)","structured message history (array of prior turns)"],"output_types":["text (assistant response)","streaming text (for real-time chat interfaces)"],"categories":["text-generation-language","memory-knowledge"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_6","uri":"capability://text.generation.language.creative.writing.and.content.generation","name":"creative writing and content generation","description":"Hermes 2 Pro generates creative content including stories, poetry, marketing copy, and other written material by learning patterns from diverse text examples in the OpenHermes dataset. The model uses transformer-based text generation to produce coherent, contextually appropriate content that follows specified styles, tones, or formats. Generation works through next-token prediction with attention to prompt specifications, enabling the model to adapt writing style, maintain narrative consistency, and follow structural requirements (e.g., sonnet format, product description length).","intents":["generate marketing copy, product descriptions, or promotional content","create story outlines, plot summaries, or narrative content","write poetry or creative text in specified styles or formats","generate blog posts, articles, or long-form content from outlines"],"best_for":["content creators and marketers needing bulk content generation","teams building writing assistance tools or content platforms","developers creating creative AI applications (storytelling games, writing assistants)"],"limitations":["Generated content may lack originality or contain clichés due to training on common patterns","No built-in fact-checking — creative content may contain false claims or hallucinations","Tone and style consistency may degrade in longer pieces (>2K tokens) due to attention dilution","No awareness of brand voice or domain-specific terminology without explicit prompt engineering"],"requires":["OpenRouter API key","Detailed prompt specifications including style, tone, length, and format requirements","Optional: examples of desired writing style for few-shot prompting"],"input_types":["text (prompts with creative specifications)","structured prompts with style, tone, and format guidelines","examples of desired writing style (few-shot)"],"output_types":["text (creative content in specified format)","streaming text (for real-time generation)"],"categories":["text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_7","uri":"capability://text.generation.language.question.answering.with.knowledge.synthesis","name":"question answering with knowledge synthesis","description":"Hermes 2 Pro answers questions by synthesizing information from the provided context or its training knowledge, using transformer attention to identify relevant information and generate coherent answers. The model processes questions and context together, attending to relevant passages and combining information across multiple sources to produce comprehensive answers. Question answering works through next-token prediction where the model learns to extract relevant facts, synthesize them, and present them in a clear, organized manner based on training examples.","intents":["answer user questions about provided documents or knowledge bases","create FAQ systems or knowledge base search interfaces","build question-answering chatbots for customer support or education","extract and synthesize information from multiple sources to answer complex questions"],"best_for":["teams building knowledge base search or FAQ systems","developers creating customer support chatbots with document context","builders prototyping question-answering applications without specialized QA infrastructure"],"limitations":["No built-in retrieval — requires external search or RAG system to identify relevant context","Answer quality depends heavily on context quality and relevance; irrelevant context may confuse the model","No source attribution — model may synthesize information without indicating which sources it came from","Hallucination risk when context doesn't contain answer; model may generate plausible-sounding but false information"],"requires":["OpenRouter API key","Question text","Relevant context or documents (provided in prompt)","Optional: external retrieval system (vector DB, search engine) to identify relevant context"],"input_types":["text (question)","text (context or documents)","structured prompts with question and context"],"output_types":["text (answer)","structured data (answer with source citations if post-processed)"],"categories":["text-generation-language","search-retrieval"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"openrouter-nousresearch-hermes-2-pro-llama-3-8b__cap_8","uri":"capability://text.generation.language.instruction.following.with.complex.multi.step.directives","name":"instruction-following with complex multi-step directives","description":"Hermes 2 Pro follows complex, multi-step instructions by parsing user directives and executing them sequentially or in parallel as appropriate. The model was explicitly trained on instruction-following examples in the OpenHermes dataset, learning to understand nuanced requirements, handle edge cases, and produce outputs that precisely match specifications. Instruction-following works through attention to instruction tokens and learned patterns that map instruction semantics to appropriate output generation, enabling the model to handle conditional logic, formatting requirements, and complex constraints.","intents":["execute complex workflows defined in natural language instructions","follow detailed specifications for output format, structure, or content","handle conditional logic or branching based on instruction parameters","process multi-part requests that require sequential or parallel execution"],"best_for":["developers building instruction-driven agents or automation systems","teams creating workflow automation tools with natural language interfaces","builders prototyping systems that require precise instruction adherence"],"limitations":["Instruction comprehension degrades with very long or ambiguous instructions (>500 tokens)","Complex conditional logic may be misunderstood; explicit formatting of conditions improves accuracy","No built-in error handling or recovery — failed instruction execution requires manual intervention","Instruction adherence varies by domain; instructions outside training distribution may be misinterpreted"],"requires":["OpenRouter API key","Clear, well-structured instructions in natural language","Optional: examples of desired output format (few-shot prompting)"],"input_types":["text (natural language instructions)","structured prompts with explicit formatting or constraint specifications"],"output_types":["text (output following instruction specifications)","structured data (if instruction specifies format)"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":25,"verified":false,"data_access_risk":"high","permissions":["OpenRouter API key or compatible OpenAI-format API endpoint","HTTP client library (curl, Python requests, JavaScript fetch, etc.)","JSON schema definition for function calling (OpenAI function_calling or similar format)","Handling for streaming or non-streaming response modes","OpenRouter API key or compatible endpoint","Code context provided in prompt (file snippets, function signatures, or full files)","Programming language specification in prompt for accurate syntax generation","Post-processing for syntax validation and linting","OpenRouter API key","Source language specification in prompt (for translation)"],"failure_modes":["8B parameter size limits reasoning depth on highly complex multi-step problems compared to 70B+ models","function calling accuracy depends on clarity of schema definition; ambiguous schemas may produce malformed JSON","context window size (likely 8K tokens based on Llama-3 8B standard) constrains conversation history and document context","no guaranteed output format validation — requires downstream JSON parsing and error handling","8B parameter size struggles with very long files (>2K lines) or complex architectural decisions requiring deep codebase understanding","no built-in awareness of project structure, dependencies, or type definitions — requires explicit context in prompt","generated code may contain logical errors or inefficiencies; always requires review and testing","no real-time incremental generation optimization — full response must be generated before streaming begins","Translation quality varies by language pair; less common language combinations may produce lower-quality output","No specialized translation memory or terminology database — may produce inconsistent terminology across documents","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.05,"quality":0.43,"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":"2026-05-03T15:20:45.777Z","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-2-pro-llama-3-8b","compare_url":"https://unfragile.ai/compare?artifact=nousresearch-hermes-2-pro-llama-3-8b"}},"signature":"bDFPkA2ehRSzW0Q+S5qrJtzlRV61IvFTXkJvY3WLBmRRVShA9VcJcNhyc9/DWwsO9WzBpo4zkIAGWvzobi/+CQ==","signedAt":"2026-06-15T17:51:19.487Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/nousresearch-hermes-2-pro-llama-3-8b","artifact":"https://unfragile.ai/nousresearch-hermes-2-pro-llama-3-8b","verify":"https://unfragile.ai/api/v1/verify?slug=nousresearch-hermes-2-pro-llama-3-8b","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"}}