Explore your specialist teams, their strengths and how their recorded performance changes over time.
“Einstein” is the visual inspiration for Exceptional, not a claim of human-equivalent intelligence. AGI cannot be unlocked by this score.
Loading your recorded assessments…
Loading execution measurements…
75 roles · capabilities below describe configured responsibilities, not verified connections.
Answer clearly and helpfully across everyday questions, explanations and conversation.
No recorded runs in this sample.
fast routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Work through hard, multi-step problems carefully, check intermediate results and state the final answer plainly.
No recorded runs in this sample.
reasoning routing · 2 maximum steps · Pro access
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Teach concepts at the learner's level with examples, checks for understanding and practice questions.
No recorded runs in this sample.
standard routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Answer questions that depend on current information by searching the web and citing the sources used.
No recorded runs in this sample.
search routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Run several searches from different angles, compare sources, resolve conflicts and write a structured, cited report.
No recorded runs in this sample.
pro routing · 8 maximum steps · Pro access
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Verify claims against reliable current sources and say clearly what is confirmed, disputed or unknown.
No recorded runs in this sample.
search routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Compare products using current prices, specs and reviews and recommend options that fit the user's needs and budget.
No recorded runs in this sample.
search routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Plan trips with itineraries, logistics and current practical details, flagging anything the user should confirm.
No recorded runs in this sample.
search routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Write, explain and fix production-quality code with tests and clear reasoning about edge cases.
No recorded runs in this sample.
code routing · 3 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Review code for correctness, security, performance and maintainability and propose concrete fixes.
No recorded runs in this sample.
code routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Help with deployment, CI/CD, containers, cloud configuration and incident debugging.
No recorded runs in this sample.
code routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Analyze tabular data, compute statistics, design charts and explain what the numbers mean.
No recorded runs in this sample.
code routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Design schemas, write and optimize queries and plan migrations safely.
No recorded runs in this sample.
code routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Design scalable, reliable systems and explain the trade-offs between approaches.
No recorded runs in this sample.
reasoning routing · 2 maximum steps · Pro access
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Solve math problems exactly, show the key steps and check the answer.
No recorded runs in this sample.
reasoning routing · 1 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Explain science accurately, distinguish established results from open questions and cite where useful.
No recorded runs in this sample.
standard routing · 3 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Draft and edit clear, natural writing in the user's voice for any format.
No recorded runs in this sample.
standard routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Write persuasive, honest marketing copy, ads, landing pages and social posts.
No recorded runs in this sample.
standard routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Draft, summarize and, with the user's confirmation, send or organize email through connected accounts.
No recorded runs in this sample.
standard routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Translate accurately while preserving tone, register and meaning, noting cultural nuances.
No recorded runs in this sample.
standard routing · 1 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Summarize long text, documents or threads into accurate key points and action items.
No recorded runs in this sample.
fast routing · 3 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Analyze markets, competition and strategy and turn them into prioritized, practical recommendations.
No recorded runs in this sample.
pro routing · 5 maximum steps · Pro access
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Explain financial concepts, model scenarios and analyze numbers; never present general information as personal financial advice.
No recorded runs in this sample.
reasoning routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Research prospects from public sources, draft outreach and update the user's CRM after confirmation.
No recorded runs in this sample.
standard routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Explain legal concepts and documents in plain language, identify jurisdiction-specific questions and recommend a licensed lawyer for advice.
No recorded runs in this sample.
reasoning routing · 4 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Help write job descriptions, interview plans, policies and performance feedback fairly and lawfully.
No recorded runs in this sample.
standard routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Find availability and, with the user's confirmation, create or change calendar events in connected calendars.
No recorded runs in this sample.
standard routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Find, read and organize files in connected drives and document tools, and draft new documents.
No recorded runs in this sample.
standard routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Plan projects, break work into tasks and, with confirmation, sync tasks to connected project tools.
No recorded runs in this sample.
standard routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Summarize and draft messages for team chat tools and post them only after the user confirms.
No recorded runs in this sample.
standard routing · 5 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Complete tasks across the user's connected apps, confirming every action that changes data or contacts other people.
No recorded runs in this sample.
standard routing · 8 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Turn ideas into detailed, original image and video prompts and art direction.
No recorded runs in this sample.
standard routing · 1 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Generate many diverse, original ideas, then help narrow them to the strongest few.
No recorded runs in this sample.
fast routing · 1 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Give accurate general health information, explain when to see a professional and never diagnose.
No recorded runs in this sample.
standard routing · 3 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Help with career decisions, job search, interviews and professional growth.
No recorded runs in this sample.
standard routing · 2 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Remember the user's stated preferences and help plan their day, errands and routines.
No recorded runs in this sample.
fast routing · 3 maximum steps
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Turn the objective into a measurable strategy with assumptions, constraints, milestones, tradeoffs and kill criteria.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Analyse evidence, market structure, competitors, demand, positioning, information gaps and confidence.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Design acquisition loops, offers, channels, experiments, retention mechanics and measurable growth systems.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Find practical paths to revenue, pricing, pipeline, conversion, distribution and commercial execution.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Create communication angles, scripts, hooks, editorial systems and channel-specific content plans.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Translate user needs into product requirements, UX flows, experiments, prioritisation and product decisions.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Design implementation architecture, code changes, tests, migrations, reliability controls and deployment plans.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Connect APIs and workflows, remove manual work, define triggers and design observable reliable automations.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Model unit economics, budgets, ROI, scenarios, cash impact, sensitivities and financial constraints.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Identify legal, regulatory, contractual and compliance issues without pretending to replace qualified counsel.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Threat-model the proposal and identify identity, data, authorization, abuse, privacy and operational security risks.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Convert the plan into sequence, owners, dependencies, checkpoints, SOPs, verification and execution cadence.
No recorded runs in this sample.
Evidence-aware multi-agent workflow
Charts use the last 30 recorded verified-outcome accuracy assessments available in this session. These are submitted assessments, not an independent intelligence certification.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Build the operating brand model: positioning, audience, promise, proof, offer, tone, visual codes, constraints and claims that are actually supported.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Map competitor hooks, offers, formats, landing-page patterns and category conventions. Extract patterns without copying creative expression.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Translate the objective into differentiated campaign hypotheses tied to audience motivation, offer structure and measurable business outcomes.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Create distinct hooks, headlines, proof structures and calls to action. Reject invented claims and vague generic copy.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Turn each hypothesis into a production-ready visual system: composition, hierarchy, typography, product treatment, scene, aspect ratio and generation prompt.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Choose the strongest available rendering workflow and exact connector/tool candidates; make missing dependencies explicit.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Challenge brand fidelity, unsupported claims, illegible layouts, platform readiness, accidental imitation and insufficient variation. Veto weak assets.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Design a controlled creative test matrix with variables, metrics, sample logic, learning objectives and stop/iterate rules.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Separate safe autonomous reads from consequential writes, route external mutations through the existing ELP approval system, and verify every claimed completion.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Own the outcome end-to-end, reconcile specialist disagreements, select the strongest concepts and return an execution-ready campaign package plus explicit next actions.
No recorded runs in this sample.
Creative mission role
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Choose the strongest available structured data source for every subtask, check freshness and provenance, and expose gaps rather than invent data.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Map regime, liquidity, breadth, trend, volatility and candidate instruments from verified market observations.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Assess price structure, trend, momentum, support/resistance, relative strength and invalidation from verified series.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Assess business quality, valuation drivers, earnings sensitivity, balance-sheet risk and fundamental catalysts when relevant.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Compare peers, competitors, sectors and benchmarks to identify leadership, divergence and correlated movement.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Identify dated market-moving news, earnings, macro events and narrative changes; reject stale or unattributed claims.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Turn hypotheses into falsifiable rules, backtests, benchmarks, transaction-cost assumptions and walk-forward validation.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Assess volatility regime and derivatives implications only when reliable options/volatility data is actually available.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Challenge every thesis, define downside cases, sizing constraints, concentration limits, invalidation and hard stop conditions; veto unsupported setups.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Combine surviving hypotheses into a diversified paper portfolio with measurable review cadence and attribution.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Detect repeated capability gaps and specify reusable skills with inputs, outputs, connectors, tests and safety boundaries.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Reconcile specialist disagreements and produce one evidence-aware operating plan with a visible agent/tool pipeline.
No recorded runs in this sample.
Research, backtesting and paper trading
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Answer parish and community questions.
No recorded runs in this sample.
Public Q&A; no external actions
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Help describe and draft a watch request.
No recorded runs in this sample.
Request drafting; inventory and pricing feeds are not verified
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Concierge assistance for the Renaissance experience.
No recorded runs in this sample.
Concept workflow; staff drafts require human review
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Sales assistance for the Renaissance experience.
No recorded runs in this sample.
Concept workflow; staff drafts require human review
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.
Marketing assistance for the Renaissance experience.
No recorded runs in this sample.
Concept workflow; staff drafts require human review
This role does not yet publish accuracy history to this directory.
A changing score shows observed task performance. It does not demonstrate model retraining or general intelligence.