OpenAI GPT-6 Family · Sol 1.2 · Released November 17, 2026
20 free GPT-6.2 Sol prompts for science research, knowledge-work documents, agentic coding, and Ultrafast batch tasks. OpenAI's most advanced Sol model — near-Astra on FrontierCode 2.0 at 1/5 the cost. No signup.
This page gives you 20 copy-paste prompts built for GPT-6.2 Sol — OpenAI's latest Sol-tier model, released November 17, 2026 alongside GPT-6.2 Astra. GPT-6.2 Sol delivers targeted improvements over GPT-6.1 Sol: a 9-point gain on OpenAI's science benchmark, +6.2 points on the GDPval v2 knowledge-work test with 22% fewer output tokens, and near-parity with GPT-6.2 Astra on FrontierCode 2.0 (within 0.5 points). Ultrafast mode — previously Astra-only in ChatGPT — now works with Sol on Pro 200 plans, enabling high-throughput batch generation.
The prompts here are built around GPT-6.2 Sol's specific new strengths: science and research tasks, structured professional documents (strategy briefs, PowerPoint decks, Excel models), near-Astra agentic coding workflows, and Ultrafast-optimised batch task designs. Each prompt includes bracketed placeholders — fill them in and paste directly into ChatGPT Work, Codex, or the API (gpt-6.2-sol).
9-point improvement on OpenAI's science benchmark at <1/3 old Astra cost
Better strategy docs, presentations, and reports with 22% fewer output tokens
First Sol model with Ultrafast for ChatGPT Pro 200 — high-speed batch tasks
Within 0.5 pts of GPT-6.2 Astra on coding benchmark at 1/5 the cost
Built for science research, knowledge-work documents, agentic coding, and Ultrafast batch tasks. Works on GPT-6.2 Sol, GPT-6.1 Sol, and Astra.
Write a structured literature review on the following research topic. Topic: [your research question or hypothesis] Scope: [date range, field, specific journals or sources to prioritise] Audience: [peer researchers / grant committee / general academic / undergraduates] Literature provided: [paste all papers, abstracts, or key passages — GPT-6.2 Sol's 1M context handles large corpora] Review structure: 1. FIELD OVERVIEW — what is known and agreed (cite sources) 2. METHODOLOGICAL LANDSCAPE — dominant methods, their strengths and common limitations 3. KEY DEBATES — where researchers disagree and why each position is held 4. GAPS AND OPEN QUESTIONS — what the field has not yet answered, ranked by importance 5. SYNTHESIS — your assessment of the most promising direction, with confidence level 6. REFERENCES — full citation list formatted to [APA / MLA / Vancouver / other] Note where evidence is thin, contested, or drawn from low-quality sources. Distinguish established fact from emerging consensus.
Perform a deep refactor of the following code. Prioritise correctness, maintainability, and performance — in that order. Language/framework: [specify] Refactoring goal: [describe what should improve — readability / performance / testability / security / architecture] Constraints: [breaking changes allowed? API surface must stay stable? minimum test coverage target?] Existing code: ``` [paste the full file or module] ``` Known issues or smells: [list anything the team already knows is wrong] Deliver: 1. ANALYSIS — the top 5 problems with the current code, each with: location, severity, and why it matters 2. REFACTORED CODE — complete, production-ready, no truncation, no 'TODO: fill in' stubs 3. CHANGE LOG — a diff-style summary: what changed and why (not how) 4. TESTS — unit tests for any logic that was unclear or bug-prone before the refactor 5. MIGRATION NOTES — if any public interface changed: before/after signature and the upgrade path
Write a strategy brief for the following business decision or initiative. Decision or initiative: [describe clearly — what is being proposed and why now] Audience: [executive team / board / investors / department heads] Data and context: [paste all relevant metrics, market data, internal reports, or competitive intelligence — GPT-6.2 Sol's 1M context handles full research dumps] Brief format (match the precision of a McKinsey memo): 1. SITUATION (2–3 sentences: what is true today, stated as fact not opinion) 2. COMPLICATION (what has changed or will change that makes action necessary) 3. RECOMMENDATION (one specific, actionable sentence — what to do) 4. SUPPORTING EVIDENCE (3 data-backed reasons — each with the source and the implication) 5. RISKS AND MITIGATIONS (top 3 risks; for each: probability, impact, mitigation) 6. RESOURCE REQUIREMENT (headcount, budget, timeline — be specific) 7. DECISION GATE (what metric or event would trigger a strategy change) Maximum 600 words. No filler. No hedged conclusions.
Build a complete, presentation-ready PowerPoint slide deck for the following topic. Topic: [describe the presentation — purpose, subject, context] Audience: [who will see this — executives / clients / investors / team / conference] Desired outcome: [what should the audience believe or do after seeing this?] Slide count: [target number] Tone: [formal executive / startup pitch / educational / technical] Data and content to include: [paste all source material — reports, metrics, talking points, previous slides, background research] For each slide, provide: - SLIDE TITLE (clear, action-oriented, ≤10 words) - KEY VISUAL (describe the chart, diagram, or image that belongs here — specific, not generic) - BULLET POINTS (3–5 max, each a complete thought in ≤12 words) - SPEAKER NOTE (2–3 sentences the presenter says — not the bullets repeated) End with: NARRATIVE ARC — a one-paragraph summary of the story the deck tells from slide 1 to the final call-to-action.
Design a financial model for the following scenario. Produce a complete specification a finance team can build in Excel or Google Sheets. Scenario: [describe the business situation — startup projection / M&A analysis / product P&L / budget vs actual] Time horizon: [monthly/quarterly/annual, and for how many periods] Key inputs: [the variables the model must accept — pricing, volume, headcount, COGS, etc.] Key outputs: [the metrics the model must calculate — revenue, margin, EBITDA, IRR, payback, cash position] Deliver: 1. MODEL STRUCTURE — tab names, what each tab contains, and how they link 2. ASSUMPTIONS TABLE — every input as a named cell with a default value and a brief rationale 3. FORMULA SPECIFICATIONS — for each key output: the calculation logic in plain English and the Excel formula 4. SENSITIVITY TABLE — the 3 inputs that swing the headline metric most, and the range to test 5. SCENARIO DEFINITIONS — Base / Upside / Downside assumptions with probability weights 6. VALIDATION CHECKS — the 5 sanity checks to add (e.g. costs never exceed revenue in a viable model)
Generate a set of research hypotheses for the following scientific question or domain. Research area: [field, sub-field, or specific phenomenon] Current understanding: [what is established and what is contested] Available methods: [what experimental or observational methods are feasible] Constraints: [ethical, budget, timeline, data-access limits] Background literature: [paste relevant papers or summaries] For each hypothesis: 1. HYPOTHESIS STATEMENT — one precise, falsifiable statement (H1 form) 2. THEORETICAL BASIS — the mechanism that would explain a positive result 3. PREDICTED OUTCOME — what a confirming experiment would show, with effect direction and size estimate 4. COMPETING EXPLANATION — the most plausible alternative hypothesis and how to distinguish between them 5. METHODOLOGICAL FIT — the strongest study design for this hypothesis given the constraints above 6. PRIORITY SCORE — why this hypothesis is worth testing ahead of the others (impact × feasibility) Generate 5 hypotheses ranked by priority.
Write a complete agentic task specification that GPT-6.2 Sol in Codex can execute end-to-end without human clarification. Task goal: [describe the software change or feature to implement] Repository context: [language, framework, relevant file paths, coding conventions] Acceptance criteria: [what must be true for the task to be complete — write as testable assertions] Out of scope: [explicitly list what the agent must NOT change] Available tools: [file read/write, shell execution, test runner, linter, git] Specification must include: 1. STARTING STATE — what the agent reads first and why 2. IMPLEMENTATION STEPS — numbered, each with: action, file(s) affected, validation check 3. SELF-VERIFICATION — the exact commands to run before declaring success 4. ROLLBACK PLAN — how the agent restores the original state if a step fails mid-way 5. EDGE CASES — the 3 scenarios the implementation must handle that aren't obvious from the goal
I am using GPT-6.2 Sol with Ultrafast mode enabled on ChatGPT Pro 200. Help me plan the following high-volume task as a sequence optimised for speed. Task: [describe what I want to produce — e.g. 50 social media posts, 100 email subject lines, rapid code review of 30 PRs, batch data classification] Input format: [how the data or content arrives — one at a time / as a batch / from a file / pasted directly] Output format: [what format the final output should take — table / list / JSON / plain text / file] Quality bar: [minimum quality acceptable for this Ultrafast batch — 'good enough to ship' / 'draft for human review' / 'must be publication-ready'] Deliver: 1. OPTIMAL PROMPT STRUCTURE — the single prompt template to run in Ultrafast mode (write it, don't describe it) 2. BATCH SIZE — how many items to process per message for maximum throughput 3. QUALITY CHECKPOINTS — when to pause Ultrafast and do a full-quality check 4. OUTPUT ASSEMBLY — how to combine the batched outputs into the final deliverable 5. TOTAL TIME ESTIMATE — rough time to complete at Ultrafast mode speed
Write a competitive intelligence report on the following market segment. Our product: [brief description — what we do, who we serve, our price point] Competitors to cover: [list 3–5 by name] Focus areas: [product features / pricing / go-to-market / funding / hiring signals / customer reviews] Intelligence sources: [paste all source material — press releases, job listings, pricing pages, review excerpts, blog posts, funding announcements — GPT-6.2 Sol's 1M context handles very large dumps] Report format: 1. EXECUTIVE SUMMARY (3 bullets: what we know, what it means for us, the one recommended action) 2. COMPETITOR PROFILES (for each: core positioning, apparent strengths, exploitable weaknesses, their most likely next move) 3. MARKET DYNAMICS (what is shifting and who benefits from the shift) 4. PRICING MAP (a table: competitor × pricing tier × key differentiator) 5. OUR BEST MOVE (one specific action in the next 30 days) 6. INTELLIGENCE GAPS (what we need to know but don't — and how to get it)
Perform a deep, senior-engineer-level code review on the following diff. Language/framework: [specify] PR description: [what this change is supposed to do] Code diff: ``` [paste the full diff] ``` Known risks the team already has: [list any] Review dimensions (cover all): 1. CORRECTNESS — logic errors that produce wrong output in edge cases. For each: line, failure scenario, fix. 2. SECURITY — injection, auth bypass, secrets exposure, unvalidated input. For each: severity, exploit path, fix. 3. PERFORMANCE — O(n²) loops, N+1 queries, blocking I/O, unbounded memory. For each: location, impact, fix. 4. AGENT SAFETY — prompt injection vectors, unintended tool calls, missing confirmation gates (if this code runs in an agent context). 5. TEST GAPS — 3 inputs most likely to expose bugs not caught by existing tests. Final verdict: APPROVE / REQUEST CHANGES / DO NOT SHIP — one-sentence rationale.
Write the following section of a scientific grant proposal. Section: [Specific Aims / Research Strategy / Significance / Innovation / Approach / Budget Justification / other] Funding body: [NIH / NSF / Wellcome Trust / ERC / DARPA / other] — match the required format Project summary: [describe the research in 2–3 sentences: the problem, the approach, the expected impact] PI background: [relevant qualifications and prior publications to cite] Preliminary data: [paste all preliminary results, figures, or key data — GPT-6.2 Sol handles large data dumps] Writing requirements: - Match the precise tone and structure of a funded proposal in this field - Use active voice; every paragraph must state a claim and support it with evidence - Significance section: lead with the clinical/societal impact, then the scientific gap - Approach section: include timeline, milestones, alternative approaches, and risk mitigations - Reviewers score Innovation separately — explicitly state what is novel Word limit: [specify, or use the section default]
Design a production-ready data pipeline for the following use case. Use case: [describe what the pipeline must do — ingest, transform, enrich, store, serve] Data sources: [list each source — type, volume per day, format, freshness requirement] Data consumers: [list each downstream system or team — their latency requirement and query pattern] Technology constraints: [must use / prefer / avoid — cloud provider, existing stack, budget] Compliance requirements: [PII handling, audit trail, data residency, retention policy] Deliver: 1. ARCHITECTURE DIAGRAM (text-based: each component → transformation → destination with latency annotations) 2. COMPONENT SPEC (for each component: technology choice, why it was chosen over alternatives, failure mode) 3. DATA CONTRACT (the schema at each stage — field names, types, nullability, example values) 4. MONITORING PLAN (5 metrics to track, the alert threshold for each, and the on-call runbook entry) 5. COST ESTIMATE (back-of-envelope: storage + compute + transfer at the stated data volume) 6. MIGRATION PLAN (how to cut over from the current system without downtime or data loss)
Synthesise the following document set into a single coherent output. Synthesis goal: [what output you need — an executive summary / a gap analysis / a merged policy document / a decision memo / a due-diligence report] Audience: [who will read the output] Documents: [paste all source documents — GPT-6.2 Sol's 1M context window handles very large document sets; paste everything] Synthesis instructions: 1. CONSOLIDATE — merge overlapping content into a single, non-repetitive account; flag where sources directly contradict each other 2. PRESERVE — do not omit any claim that appears in only one source; mark low-confidence or single-source claims explicitly 3. STRUCTURE — organise by theme, not by source document; use headers 4. CITE — after each significant claim, note which source(s) support it (e.g. [Doc 2]) 5. GAPS — at the end, list what the documents collectively do not address that would be needed for [the synthesis goal] Output format: [report / memo / bullet list / structured JSON — specify]
Write a market research report for the following product, market, or investment thesis. Subject: [what you are researching — a product category, a company, a geographic market, an emerging trend] Research question: [the core question this report must answer] Audience: [investors / product team / sales / leadership / external clients] Research material: [paste all source material — industry reports, news, company filings, interview notes, survey data — GPT-6.2 Sol handles full research dumps] Report structure: 1. EXECUTIVE SUMMARY (3 bullets: key finding, market opportunity, recommended action) 2. MARKET SIZE AND GROWTH (TAM / SAM / SOM with source citations and growth rate) 3. CUSTOMER SEGMENTATION (3–5 buyer personas: who they are, their pain, their current solution, their switching cost) 4. COMPETITIVE LANDSCAPE (table: top 5 players × funding / revenue / growth / positioning) 5. TREND ANALYSIS (3 tailwinds and 2 headwinds, each with evidence and time horizon) 6. OPPORTUNITY MAP (where is the unmet need; why now) 7. RISKS (regulatory, technological, adoption — ranked by probability × impact)
Write production-ready technical documentation for the following system or API. System or API to document: [name and brief description] Documentation type: [API reference / Getting Started guide / Integration tutorial / Architecture overview / Troubleshooting guide] Audience: [developers / platform engineers / technical PMs / data scientists] Tech stack: [language, framework, relevant dependencies] System details: [paste the code, schema, OpenAPI spec, architecture diagrams, or existing rough notes — GPT-6.2 Sol handles large specs] Documentation must include: 1. OVERVIEW — what this system does, when to use it, and what it replaces or integrates with 2. QUICKSTART — the minimum working example (copy-paste, fully runnable, annotated) 3. REFERENCE — every endpoint / method / config option: parameters, types, defaults, example request/response 4. ERROR CATALOGUE — every error code, its cause, and the resolution 5. BEST PRACTICES — 5 specific dos and don'ts based on the system's actual design 6. CHANGELOG HOOK — a placeholder section marking where version history should live
Design a rigorous experiment protocol for the following research question. Research question: [state the question precisely — what are you measuring and why] Hypothesis: [the specific, falsifiable prediction] Experimental context: [field, lab vs. field, human vs. animal vs. in-vitro, duration] Available resources: [instruments, sample sizes, budget, timeline] Constraints: [ethical approval required, blinding requirements, pre-registration requirement] Protocol must include: 1. STUDY DESIGN — type (RCT / observational / longitudinal / A/B / other), rationale for the choice 2. SAMPLE SPECIFICATION — inclusion/exclusion criteria, minimum N with power calculation, randomisation method 3. INTERVENTION OR TREATMENT — exact description of what each group receives; control condition definition 4. MEASUREMENT PLAN — primary outcome (the one that answers the hypothesis), secondary outcomes, measurement instrument and timing 5. STATISTICAL ANALYSIS PLAN — test(s) to use, significance threshold (α), how to handle missing data 6. BIAS CONTROLS — blinding protocol, allocation concealment, confounders to measure and adjust for 7. STOPPING RULES — conditions under which the experiment should stop early (safety, futility, overwhelming efficacy)
Write a complete Product Requirements Document (PRD) for the following feature or product. Feature/product: [describe what is being built] Problem it solves: [the user pain and the business opportunity] Target users: [primary persona — their context, goals, and current workaround] Success metrics: [how you will know this feature succeeded — 1–3 measurable outcomes] Constraints: [technical debt, platform limitations, compliance, timeline, team size] Background material: [paste user research, stakeholder interviews, prior spec drafts, design mocks descriptions, competitive examples — GPT-6.2 Sol handles large context] PRD sections: 1. PROBLEM STATEMENT (2–3 sentences: the pain, its frequency, its cost to the user) 2. GOALS AND NON-GOALS (what this version does and explicitly does NOT include) 3. USER STORIES (5–8 stories in the format: As a [persona], I want to [action] so that [benefit]; acceptance criteria for each) 4. FUNCTIONAL REQUIREMENTS (numbered list; each must be testable as pass/fail) 5. EDGE CASES AND ERROR STATES (5 scenarios engineers must handle) 6. OPEN QUESTIONS (blockers that need a decision before engineering starts)
Design a multi-agent workflow to automate the following complex task. Task: [describe the end-to-end goal] Data inputs: [what information the workflow starts with] Expected outputs: [what the workflow must produce] Enterprise systems involved: [CRM / ERP / data warehouse / communication tools / APIs — list all] Human oversight required: [which decisions must have a human in the loop, and why] Deliver a workflow specification: 1. AGENT MAP — list each agent, its role, its tools, and its authority boundary (what it can and cannot do) 2. ORCHESTRATION LOGIC — the control flow: triggers, branching conditions, parallel paths, merge points 3. DATA HANDOFFS — what each agent passes to the next and in what format (schema required for each handoff) 4. FAILURE MODES — the 5 most likely failure scenarios; for each: how it is detected and what the recovery action is 5. AUDIT TRAIL — what each agent logs and where (required for compliance and debugging) 6. COST MODEL — estimated token usage per run; the most expensive step and how to optimise it 7. TEST PLAN — 3 end-to-end test cases with inputs, expected outputs, and pass/fail criteria
Write a production-ready system prompt for a GPT-6.2 Sol agent deployed in [ChatGPT Work / Codex / the API]. Agent role: [describe what the agent does — e.g. 'senior code reviewer for a fintech API', 'scientific literature analyst', 'enterprise document drafter'] Audience who will interact with it: [developers / end users / other agents] Tone and style: [formal / technical / concise / collaborative — be specific] Tools available: [list all — web search, code execution, file access, APIs, other agents] Constraints: - [constraint 1 — e.g. 'never make recommendations about financial products'] - [constraint 2 — e.g. 'always cite sources for factual claims'] - [constraint 3] Output format for responses: [structured / prose / JSON / markdown — specify] System prompt must include: 1. ROLE DEFINITION (one paragraph — who the agent is, what it does, what it doesn't do) 2. OPERATING RULES (numbered; each rule is specific and testable, not vague) 3. TOOL USE PROTOCOL (when and how to use each tool — the order and the validation step) 4. RESPONSE FORMAT SPEC (the exact structure of every response — headers, fields, length target) 5. ESCALATION CONDITIONS (when to stop and ask the user rather than proceeding) 6. CHAIN-OF-THOUGHT INSTRUCTION (how visible reasoning should be — hidden / shown / summarised)
Prepare a complete executive briefing package for the following decision. Decision: [state precisely what must be decided] Decision deadline: [when] Decision makers: [title/role of each person who must approve] Stakeholders affected: [teams, customers, regulators impacted by the decision] Background material: [paste all context — data, reports, models, prior discussion — GPT-6.2 Sol handles full research packages] Briefing package (one page per section): 1. ONE-PAGE SUMMARY — the decision, the recommendation, and the three most important facts 2. SITUATION ANALYSIS — what is true today (data-backed, cited) 3. OPTIONS ANALYSIS — 3 options in a table: description / pros / cons / cost / timeline / risk 4. RECOMMENDATION — the preferred option, stated directly, with the three reasons and the one strongest counterargument 5. FINANCIAL IMPACT — best / base / downside at 1 year and 3 years 6. IMPLEMENTATION ROADMAP — 90-day action plan if the recommendation is approved: owner, action, deadline for each step 7. APPENDIX — raw data, model assumptions, sources
How GPT-6.2 Sol compares for science, knowledge-work, and agentic coding.
| Model | Best For | Coding | Science | Cost |
|---|---|---|---|---|
| GPT-6.2 Sol ✓ New | Science, knowledge-work docs, near-Astra coding, Ultrafast batch | Near-Astra | +9 pts | Medium ($2/$10) |
| GPT-6.2 Astra ✓ New | Maximum science, frontier reasoning, computer use | Maximum | Top | High |
| GPT-6.1 Sol | Agentic coding, multi-agent orchestration, long-context | Near-Astra | Good | Medium ($2/$10) |
| Claude Fable 5.2 | Nuanced reasoning, instruction adherence, coding | High | High | Medium |
| Gemini 4 Carbon | Long-context multimodal, Google ecosystem integration | High | High | Medium |
GPT-6.2 Sol is OpenAI's latest Sol-tier model, released on November 17, 2026 alongside GPT-6.2 Astra. It improves on GPT-6.1 Sol with stronger science and knowledge-work performance: +9 percentage points on OpenAI's science benchmark at less than one-third the cost of old GPT-6 Astra, and +6.2 percentage points on the GDPval v2 knowledge-work benchmark using 22% fewer output tokens. It also introduces Ultrafast mode for ChatGPT Pro 200 subscribers for the first time, enabling high-speed batch tasks. It is priced at $2.00/$10.00 per million tokens — identical to GPT-6.1 Sol.
GPT-6.2 Sol improves on GPT-6.1 Sol in two focused areas. First, science and research tasks: OpenAI reports a 9-point improvement on its science benchmark, making Sol competitive with specialist research tools. Second, knowledge-work document quality: the GDPval v2 benchmark shows +6.2 points with 22% fewer output tokens — meaning GPT-6.2 Sol writes more precise strategy memos, PowerPoint decks, and Excel models with less verbose output. It also comes within 0.5 percentage points of GPT-6.2 Astra on the FrontierCode 2.0 coding benchmark. Ultrafast mode, previously Astra-only in ChatGPT, now works with Sol on Pro 200 plans.
GPT-6.2 Sol is priced at $2.00 per million input tokens and $10.00 per million output tokens via the OpenAI API (gpt-6.2-sol). These rates are unchanged from GPT-6.1 Sol. Cached input is priced at $0.10 per million tokens — especially valuable for agentic workflows where the same long system prompt or codebase context is reused across many calls. In ChatGPT, it is available on Plus, Pro, Business, Enterprise, and Edu plans. Check OpenAI's pricing page for current rates.
Yes, if your workload involves science research, knowledge-work documents, or near-Astra coding. The key upgrades are the science benchmark improvement (+9 pts), better structured document output (PowerPoint, Excel, strategy briefs via GDPval +6.2 pts), and near-Astra parity on FrontierCode 2.0 (within 0.5 pts) — all at the same $2/$10 price. Ultrafast mode is new for Sol users on Pro 200 plans. If your usage is purely general conversational or basic coding, the improvement may be marginal. For agentic coding, science, and professional document generation, it is a meaningful step up.
Ultrafast mode is a high-throughput generation setting available to ChatGPT Pro 200 subscribers. When enabled, GPT-6.2 Sol generates output significantly faster than standard mode — trading some depth for speed, making it ideal for batch tasks: generating 50+ social posts, mass email drafting, rapid code annotation, batch data classification, or any high-volume task where speed matters more than maximum quality. GPT-6.2 Sol is the first Sol-tier model to support Ultrafast mode in ChatGPT; previously it was only available with Astra. It is not available in the standard API — it is a ChatGPT interface feature.
GPT-6.2 Sol is available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users, and via the OpenAI API as the model identifier gpt-6.2-sol. It was released on November 17, 2026. In the model selector, it appears above GPT-6.1 Sol and below GPT-6.2 Astra. Ultrafast mode for GPT-6.2 Sol in ChatGPT is rolling out to Pro 200 subscribers. It supports the same 1M+ token context window as GPT-6.1 Sol.
Previous Sol — agentic coding and multi-agent orchestration
Frontier tasks, computer use, max reasoning — 20 prompts
The original Sol — complex coding and research
Anthropic's fast agentic model — 20 copy-paste prompts
Anthropic's flagship — 20 copy-paste prompts
Google's latest — long-context multimodal prompts
More free generators in this collection — no signup, unlimited use.
50 free copy-paste Claude prompts for writing, coding, research and Claude Code — each tagged with the best Claude model.
20 copy-paste prompts for Meta Muse Wonder 1.0 — frontier multimodal AI with 1M context, text/image/video/audio input.
20 copy-paste prompts for Mistral Large 4 — cybersecurity, finance, engineering and multimodal analysis.
20 copy-paste prompts for Claude Haiku 5.5 — high-volume classification, summarization and batch tasks at $0.10/M tokens.
20 copy-paste prompts for Claude Sonnet 5.5 by effort level — fast agentic coding and visual tasks.
20 free prompts for Claude Opus 5.5 — deep reasoning for long documents and audits.