APEX-Agents benchmark
Agent benchmark tasks for real professional work.
Browse APEX-Agents benchmark tasks by real-world professional work category, from DCF valuation to legal contract review and market sizing, with public agent traces for every task.
480Tasks
7Published models
171Primary passes
Published model set
This public benchmark currently includes Gemini 3 Flash, Gemini 3.1 Pro, GPT-5.4, gpt-5.4-mini, gpt-5.4-nano, gpt-5.5, Kimi K2. Model counts reflect completed, shareable runs in this catalog.
Work categories
Browse APEX-Agents by buyer-recognizable workflow, not internal world ID.
AI Agents for Excel Financial ModelingSpreadsheet automation AIEdit sheets, recalc formulas, sensitivity tablesAI Agents for Financial Statement AnalysisSEC filings AIEBITDA, SBC addbacks, lease liabilities, EV calculationAI Agents for M&A Legal Due DiligenceM&A diligence automationHSR filings, transaction documents, deal risk allocationAI Agents for Document DraftingDocument automation AILegal letters, memos, compliance docs, Word outputsAI Agents for Pricing StrategyPricing analytics AISaaS packaging, usage-based pricing, ARR impactAI Agents for Board and Executive MemosKnowledge-work AIConcise legal, finance, and consulting written answersAI Agents for DCF ValuationFinancial modeling AIWACC, terminal value, implied share price, FCF sensitivityAI Agents for Take-Private AnalysisPE transaction AIPublic-to-private deals, offer price, shareholder valueAI Agents for Legal Contract ReviewContract review automationSupply agreements, leases, JV agreements, tariff clausesAI Agents for Merger ModelsM&A modeling AIAccretion/dilution, exchange ratio, pro forma EPSAI Agents for Portfolio StrategyStrategy consulting AIProject divestment, reallocation, valuation scenariosAI Agents for Litigation Risk AssessmentLegal memo automationLikelihood of success, motion to dismiss, claims analysis
Browse all categories across benchmarks →
All tasks
One task hub per benchmark task, with published model runs and public traces.
- Analyze Effect of Anti-SLAPP Suit Against BlueLNG on Pre-Suit Discovery(task_4d143fb2608745848a8f1096667996be)2/4 · FailPublic trace →
- Change in Competitor SG&A Spend by Category_Task02_SC(task_449e96fd1efb4a7f9ecbb65e82fbd6c1)0/7 · FailPublic trace →
- Determine Outcome of Shipyard Fire and Delay(task_8ec48c4dfa5e4f06b8bac76409c74d83)3/6 · FailPublic trace →
- Reconcile Compensation, Fee, and Cost Allocation Provisions Across All JV Documents(task_43ab089534f243eba52ee023a0508ac2)1/1 · PassPublic trace →
- World 112-1 | Task 2 - Depreciation Reduction for plants (JS)(task_6b27cc3ab9da428eaa9daa9f5100882b)6/7 · FailPublic trace →