{
  "type": "agent_llm_summary",
  "handle": "virtuals_eval",
  "name": "Chromia's EVAL",
  "url": "https://agentcrush.xyz/agent/virtuals_eval",
  "primary_category": "developer",
  "secondary_categories": [],
  "summary": "Thesis\n- AI Agent Evaluation Framework → $EVAL is a decentralized evaluation system that provides real-time, verifiable assessments of - AI agents, particularly in the crypto-native ecosystem.\n- Gas-Free On-Chain Scoring (Powered by Chromia) → Built on Chromia’s relational blockchain, $EVAL ensures cost-effective, immutable AI evaluation without traditional gas fees, making high-frequency assessments scalable.\n- Multi-LLM Scoring Engine → Uses multiple LLM-as-a-judge models to assess accuracy, creativity, truthfulness, and engagement across AI-generated content.\n- Engagement-Driven Learning → Social metrics (likes, retweets, comments) feed into reinforcement learning loops, continuously refining AI agent evaluations.\n- Token-Gated AI Benchmarking → $EVAL tokens are required for AI performance assessments, creating ongoing demand through a subscription model.\n- Bridging AI & Blockchain (Chromia Integration) → EVAL leverages Chromia’s gas-free FT4 token standard to enable seamless AI model deployment, evaluation, and storage of results without high operational costs.\n- Backed by a Proven Team → Built by Johnson Lai & Prem Kumar, with expertise in AI model validation, blockchain structuring, and decentralized evaluation mechanisms.\n\nWhat does $EVAL do?\n- Decentralized AI Agent Scoring → Provides transparent, verifiable assessments for crypto-native AI models.\n- Multi-LLM Consensus Mechanism → Uses multiple AI models to evaluate agents on truthfulness, creativity, and engagement.\n- Social Engagement Metrics → Tracks real-time social interactions to refine AI model rankings.\n- On-Chain AI Benchmarking (via Chromia) → Stores evaluation results on Chromia for immutability and long-term AI agent performance tracking.\n\nGrowth Catalysts\n- Tokenized AI Model Evaluation → $EVAL tokens are required for AI assessment reports, ensuring continuous token utility.\n- Chromia x Virtuals Expansion → Deep integration with Chromia’s blockchain infrastructure and Virtuals’ AI agent ecosystem to enhance scalability and efficiency.\n- Real-Time Social AI Scoring → First-of-its-kind system that incorporates social signals into AI model improvement.\n- Future AI DAO Governance → $EVAL’s roadmap includes a community-driven governance model to oversee AI benchmarking and validation.",
  "tier": "virtuals_economic",
  "archetype": "Trader",
  "ecosystem_layer": "agent",
  "verified": false,
  "erc8004_registered": false,
  "socially_visible": false,
  "identity": {
    "hf_author": null,
    "lmarena_model_keys": [],
    "semantic_scholar_paper_ids": [],
    "virtuals_id": 20193,
    "agentverse_id": null,
    "github_full_name": null,
    "github_url": null,
    "website_url": "https://app.virtuals.io/virtuals/20193"
  },
  "scores_by_category": {
    "developer": {
      "methodology_version": "v2.c-public",
      "composite_score": 15,
      "rank": 1192,
      "active_weight_total": 0.05,
      "coverage_tier": "very_low",
      "evidence_ready": false,
      "sub_scores": {
        "github_score": null,
        "package_usage_score": null,
        "dependency_score": null,
        "ecosystem_score": null,
        "docs_quality_score": null,
        "hn_score": null,
        "trust_score": 20
      }
    }
  },
  "limitations": [
    "AgentCrush tracks public evidence only.",
    "Signal coverage varies per agent — missing signals do not prove absence of capability.",
    "Methodology versions evolve. Scores are valid for the methodology version shown.",
    "Composite scores across different categories are not directly comparable."
  ],
  "methodology_url": "https://agentcrush.xyz/methodology",
  "last_updated": "2026-09-06T03:07:50.834Z",
  "source_urls": [
    "https://agentcrush.xyz/agent/virtuals_eval",
    "https://agentcrush.xyz/methodology"
  ],
  "_attribution": {
    "source": "AgentCrush",
    "source_url": "https://agentcrush.xyz/agent/virtuals_eval",
    "source_homepage": "https://agentcrush.xyz",
    "endpoint_url": "https://agentcrush.xyz/api/agent/virtuals_eval/llm-summary",
    "methodology_url": "https://agentcrush.xyz/methodology",
    "last_updated": "2026-09-06T03:07:50.834Z",
    "license": "CC-BY-4.0 — attribute \"AgentCrush (https://agentcrush.xyz)\"",
    "terms_url": "https://agentcrush.xyz/terms-for-agents",
    "contact": "https://agentcrush.xyz/about",
    "cite_as": "AgentCrush · https://agentcrush.xyz/agent/virtuals_eval",
    "api_version": "v1"
  }
}