# Personal AI agents compared Canonical page: https://benchlm.ai/ai-agents Sources checked: 2026-09-30 Personal AI agents differ in who runs the computer, chooses the model, and pays for usage. OpenClaw and self-hosted Hermes give you provider control. Dots, Grok Bot, Claude, and Manus manage more of the runtime, with different allowances and access rules. Compare those choices before granting access to your files or accounts. We checked public documentation for these six products on September 30, 2026. We have not installed them, checked out paid plans, or run matched tasks. The table compares documented controls and billing, with unknowns kept visible. ## Compare personal AI agents side by side Personal agent control and billing ledger. Six products in alphabetical order. No performance ranking or hands-on test. BYOK means connecting your own model-provider API key. | Decision | [Claude](https://benchlm.ai/ai-agents#claude) | [Dots](https://benchlm.ai/ai-agents#dots) | [Grok Bot](https://benchlm.ai/ai-agents#grok-bot) | [Hermes Agent](https://benchlm.ai/ai-agents#hermes-agent) | [Manus](https://benchlm.ai/ai-agents#manus) | [OpenClaw](https://benchlm.ai/ai-agents#openclaw) | | --- | --- | --- | --- | --- | --- | --- | | Hosting | **Documented.** Cloud tasks; local files and apps need Claude Desktop open.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Documented.** Own cloud computer; local computer access is optional and starts off.
Sources: [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Documented.** Cursor-hosted cloud computer. Optional local execution has separate controls.
Sources: [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security)
Checked 2026-09-30. | **Documented.** Self-hosted runtime, or optional Hermes Cloud instance.
Sources: [Hermes Agent licensing and hosting](https://hermes-agent.nousresearch.com/); [Hermes Cloud billing](https://portal.nousresearch.com/cloud)
Checked 2026-09-30. | **Documented.** Managed agent service; authorized local computer use and a paid Cloud Computer add-on are distinct options.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0)
Checked 2026-09-30. | **Documented.** Gateway on your own infrastructure: local machine or hosted server.
Sources: [OpenClaw runtime and trust boundary](https://docs.openclaw.ai/help/faq/what-is-openclaw)
Checked 2026-09-30. | | Model choice | **Documented.** A model picker remains in the unified experience.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Unknown.** GPT-6 Astra is documented. A user model picker is undocumented.
Sources: [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/)
Checked 2026-09-30. | **Documented.** Cursor selects models and failovers. No customer model picker or fixed vendor set.
Sources: [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security)
Checked 2026-09-30. | **Documented.** Configure providers and models, including compatible Ollama and vLLM endpoints.
Sources: [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers)
Checked 2026-09-30. | **Unknown.** Current foundation-model picker and bundled model roster are unverified.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0)
Checked 2026-09-30. | **Documented.** Choose a provider/model; custom and local-runtime routes are documented.
Sources: [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [OpenClaw custom and local providers](https://docs.openclaw.ai/concepts/model-providers/custom-providers)
Checked 2026-09-30. | | BYOK and authentication | **Unknown.** Claude account. General third-party model BYOK is undocumented; own-cloud marketing does not establish it.
Sources: [Claude Cowork product](https://claude.com/product/cowork)
Checked 2026-09-30. | **Unknown.** Eligible ChatGPT account. User-supplied model API keys are undocumented.
Sources: [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Unknown.** Cursor account; eligible SuperGrok or X subscription links grant usage. Model BYOK is undocumented.
Sources: [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans); [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security)
Checked 2026-09-30. | **Documented.** Provider keys, documented subscription OAuth routes, or compatible local endpoints. Each route has its own eligibility.
Sources: [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers)
Checked 2026-09-30. | **Unknown.** Manus account. User-supplied model API keys are undocumented in the checked sources.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0); [Manus credit plans](https://manus.im/pricing)
Checked 2026-09-30. | **Documented.** Provider API keys and documented subscription routes; local endpoints depend on server authentication.
Sources: [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [OpenClaw custom and local providers](https://docs.openclaw.ai/concepts/model-providers/custom-providers)
Checked 2026-09-30. | | Plan fee and included usage | **Documented.** US Pro: USD 20 per individual subscription/month. Task work consumes plan usage; regional prices differ.
Sources: [Claude Pro pricing](https://support.claude.com/en/articles/8325606-what-is-the-pro-plan); [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Documented.** Pro 100/200/500: USD 100/200/500 per individual subscription/month. Business Premium: USD 125 per seat/month, or USD 100 per seat/month billed annually; at least two total Business seats. First dot included; deeper-work allowance applies.
Sources: [ChatGPT Pro tiers](https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers); [ChatGPT Business seats](https://help.openai.com/en/articles/8792828-chatgpt-business-overview); [OpenAI Business pricing](https://openai.com/business/pricing/); [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/)
Checked 2026-09-30. | **Unknown.** Eligible Cursor plans or linked subscriptions include usage. Extra usage can be on-demand; exact numeric allowance is unknown.
Sources: [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans)
Checked 2026-09-30. | **Documented.** MIT software has no license fee. Optional Portal plans include credits; public currency code is unverified, so numeric prices are withheld.
Sources: [Hermes Agent licensing and hosting](https://hermes-agent.nousresearch.com/); [Nous Portal plans](https://portal.nousresearch.com/manage-subscription)
Checked 2026-09-30. | **Unknown.** Credit plans. Current plan amounts and selected allowances are unverified; review the billing interval and credit quantity before checkout.
Sources: [Manus credit plans](https://manus.im/pricing); [Manus task-credit billing](https://help.manus.im/en/articles/11711097-what-are-the-rules-for-credits-consumption-and-how-can-i-obtain-them)
Checked 2026-09-30. | **Unknown.** No first-party paid tier established in these sources. Provider subscriptions or usage charges are separate.
Sources: [OpenClaw runtime and trust boundary](https://docs.openclaw.ai/help/faq/what-is-openclaw); [OpenClaw usage and costs](https://docs.openclaw.ai/reference/api-usage-costs)
Checked 2026-09-30. | | Inference charges | **Documented.** Tasks use subscription allowances. Generic API and Console use is billed separately.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude); [Claude subscription and API billing](https://support.claude.com/en/articles/9876003-i-have-a-paid-claude-subscription-pro-max-team-or-enterprise-plans-why-do-i-have-to-pay-separately-to-use-the-claude-api-and-console)
Checked 2026-09-30. | **Documented.** Dot conversations do not consume ChatGPT limits; delegated Codex and ChatGPT Work tasks consume their usual usage.
Sources: [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/)
Checked 2026-09-30. | **Unknown.** Serving model affects billing. On-demand usage is billed through Cursor; Bot token rates are unverified.
Sources: [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security); [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans)
Checked 2026-09-30. | **Documented.** BYOK bills the provider. Native Anthropic OAuth uses Max extra credits, not base allowance; Portal uses its credit balance.
Sources: [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers)
Checked 2026-09-30. | **Unknown.** A separate foundation-model inference bill is not established for the consumer product. Task-credit cost per job is untested.
Sources: [Manus credit plans](https://manus.im/pricing); [Manus task-credit billing](https://help.manus.im/en/articles/11711097-what-are-the-rules-for-credits-consumption-and-how-can-i-obtain-them)
Checked 2026-09-30. | **Documented.** Configured provider and authentication route determine usage charges. Local cost estimates are not provider invoices.
Sources: [OpenClaw usage and costs](https://docs.openclaw.ai/reference/api-usage-costs)
Checked 2026-09-30. | | Tool charges | **Unknown.** Task work uses plan limits. A complete third-party tool bill is unknown.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Documented.** Delegated work uses its normal allowance. Texting may incur message or data charges.
Sources: [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/); [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Unknown.** A complete separate tool-charge schedule is not established.
Sources: [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans)
Checked 2026-09-30. | **Documented.** Own tool-provider charges apply by configuration. Nous hosted tools use credits, separate from Cloud hosting.
Sources: [Hermes Agent licensing and hosting](https://hermes-agent.nousresearch.com/); [Hermes Cloud billing](https://portal.nousresearch.com/cloud)
Checked 2026-09-30. | **Unknown.** A complete separate tool-charge schedule is not established.
Sources: [Manus credit plans](https://manus.im/pricing); [Manus task-credit billing](https://help.manus.im/en/articles/11711097-what-are-the-rules-for-credits-consumption-and-how-can-i-obtain-them)
Checked 2026-09-30. | **Documented.** Optional search, browser, and other service calls can add their own usage charges.
Sources: [OpenClaw usage and costs](https://docs.openclaw.ai/reference/api-usage-costs)
Checked 2026-09-30. | | Host and hardware charges | **Unknown.** Local tasks require your device. No separate cloud-hosting charge is established here.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Unknown.** Cloud computer is part of the product; local access uses your device. Future extra-dot prices are unknown.
Sources: [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/); [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Unknown.** Hosted computer is the supported deployment. A separate host charge is not established.
Sources: [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security); [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans)
Checked 2026-09-30. | **Documented.** Own hardware or host bill. Hermes Cloud charges uptime plus storage; stopped instances still incur storage. Inference and tools are additional.
Sources: [Hermes Cloud billing](https://portal.nousresearch.com/cloud)
Checked 2026-09-30. | **Unknown.** Cloud Computer is a paid add-on. Its current amount is unknown. Local computer use needs your device.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0)
Checked 2026-09-30. | **Documented.** You pay for the chosen server or hardware, storage, power, and upkeep. No common monthly total.
Sources: [OpenClaw runtime and trust boundary](https://docs.openclaw.ai/help/faq/what-is-openclaw); [OpenClaw custom and local providers](https://docs.openclaw.ai/concepts/model-providers/custom-providers)
Checked 2026-09-30. | | Permissions | **Documented.** Connected apps and selected folders; Manual is the default action setting, with Auto optional.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Documented.** App permissions and Custom Rules control actions. Disconnecting an app does not erase previously collected information.
Sources: [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Documented.** One user’s Bots share computer access. Local commands ask for approval by default; configurable policies can allow more.
Sources: [Grok Bot computer and access FAQ](https://docs.x.ai/grok-bot/faq); [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security)
Checked 2026-09-30. | **Documented.** Configured user allowlists, dangerous-command approval, and optional file/container restrictions. Their behavior depends on settings.
Sources: [Hermes permissions and approvals](https://hermes-agent.nousresearch.com/docs/user-guide/security)
Checked 2026-09-30. | **Documented.** Local computer use requires a connected, authorized session for approved files, browser, and apps.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0)
Checked 2026-09-30. | **Documented.** Tool policy, approvals, and sandbox settings control access. A shared Gateway is one trust domain.
Sources: [OpenClaw runtime and trust boundary](https://docs.openclaw.ai/help/faq/what-is-openclaw); [OpenClaw tool and agent permissions](https://docs.openclaw.ai/gateway/security/tool-permissions)
Checked 2026-09-30. | | Account and region access | **Documented.** Unified experience rolls out to Pro and Max on web, desktop, and mobile. Account timing varies.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude)
Checked 2026-09-30. | **Documented.** Gradual rollout: Pro outside EEA, Switzerland, and UK; Business Premium in supported ChatGPT regions; Enterprise beta needs admin enablement. Create on desktop.
Sources: [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Unknown.** Paid individual Cursor plans and Teams; eligible individual subscription links. Enterprise enablement is account-dependent; geographic eligibility is unverified.
Sources: [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans)
Checked 2026-09-30. | **Documented.** Self-hosted route needs a configured provider. Portal and Cloud need a Nous account; model and region availability depend on the route.
Sources: [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers); [Hermes Cloud billing](https://portal.nousresearch.com/cloud)
Checked 2026-09-30. | **Unknown.** Manus 2.0 is announced on web, desktop, and mobile. Detailed geographic and account eligibility is unverified.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0)
Checked 2026-09-30. | **Unknown.** Provider accounts and connected services set their own access requirements. No universal regional entitlement is established.
Sources: [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins)
Checked 2026-09-30. | | Consequential unknowns | **Unknown.** Account rollout, general BYOK, and detailed own-cloud eligibility and billing.
Sources: [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude); [Claude Cowork product](https://claude.com/product/cowork)
Checked 2026-09-30. | **Unknown.** Model picker, BYOK, exact account deeper-work allowance, and additional-dot prices.
Sources: [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/); [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot)
Checked 2026-09-30. | **Unknown.** Exact model routing and numeric usage allowances remain unknown. Linked grants do not stack; the FAQ says the larger allowance applies. Actual account charges are untested.
Sources: [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans); [Grok Bot computer and access FAQ](https://docs.x.ai/grok-bot/faq)
Checked 2026-09-30. | **Unknown.** Portal currency code, actual model/tool debits, account quotas, and the complete recurring bill.
Sources: [Nous Portal plans](https://portal.nousresearch.com/manage-subscription); [Hermes Cloud billing](https://portal.nousresearch.com/cloud)
Checked 2026-09-30. | **Unknown.** Current models, BYOK, plan/credit selections, annual invoice totals, and Cloud Computer add-on amount.
Sources: [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0); [Manus credit plans](https://manus.im/pricing)
Checked 2026-09-30. | **Unknown.** Exact provider entitlement, complete job cost, and reliability of your chosen model/tool configuration.
Sources: [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [OpenClaw usage and costs](https://docs.openclaw.ai/reference/api-usage-costs)
Checked 2026-09-30. | | Guide and test coverage | **Not tested.** No dedicated setup guide. Hands-on not run.
Sources: [Coverage note](https://benchlm.ai/ai-agents)
Checked 2026-09-30. | **Not tested.** Dots guide. Hands-on not run.
Sources: [Dots guide](https://benchlm.ai/blog/posts/openai-dots-guide)
Checked 2026-09-30. | **Not tested.** Published product overview. Hands-on not run.
Sources: [Product overview](https://benchlm.ai/blog/posts/we-measure-the-session-they-sell-the-week)
Checked 2026-09-30. | **Not tested.** Published product overview. Setup and billing guides planned; hands-on not run.
Sources: [Product overview](https://benchlm.ai/blog/posts/we-measure-the-session-they-sell-the-week)
Checked 2026-09-30. | **Not tested.** No dedicated setup guide. Hands-on not run.
Sources: [Coverage note](https://benchlm.ai/ai-agents)
Checked 2026-09-30. | **Not tested.** Published product overview. Setup and billing guides planned; hands-on not run.
Sources: [Product overview](https://benchlm.ai/blog/posts/we-measure-the-session-they-sell-the-week)
Checked 2026-09-30. | | Primary sources and read dates | [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude) (read 2026-09-30); [Claude Cowork product](https://claude.com/product/cowork) (read 2026-09-30); [Claude Pro pricing](https://support.claude.com/en/articles/8325606-what-is-the-pro-plan) (read 2026-09-30); [Claude subscription and API billing](https://support.claude.com/en/articles/9876003-i-have-a-paid-claude-subscription-pro-max-team-or-enterprise-plans-why-do-i-have-to-pay-separately-to-use-the-claude-api-and-console) (read 2026-09-30) | [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot) (read 2026-09-30); [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/) (read 2026-09-30); [ChatGPT Pro tiers](https://help.openai.com/en/articles/9793128-about-chatgpt-pro-tiers) (read 2026-09-30); [ChatGPT Business seats](https://help.openai.com/en/articles/8792828-chatgpt-business-overview) (read 2026-09-30); [OpenAI Business pricing](https://openai.com/business/pricing/) (read 2026-09-30) | [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security) (read 2026-09-30); [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans) (read 2026-09-30); [Grok Bot computer and access FAQ](https://docs.x.ai/grok-bot/faq) (read 2026-09-30) | [Hermes Agent licensing and hosting](https://hermes-agent.nousresearch.com/) (read 2026-09-30); [Hermes Cloud billing](https://portal.nousresearch.com/cloud) (read 2026-09-30); [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers) (read 2026-09-30); [Nous Portal plans](https://portal.nousresearch.com/manage-subscription) (read 2026-09-30); [Hermes permissions and approvals](https://hermes-agent.nousresearch.com/docs/user-guide/security) (read 2026-09-30) | [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0) (read 2026-09-30); [Manus credit plans](https://manus.im/pricing) (read 2026-09-30); [Manus task-credit billing](https://help.manus.im/en/articles/11711097-what-are-the-rules-for-credits-consumption-and-how-can-i-obtain-them) (read 2026-09-30) | [OpenClaw runtime and trust boundary](https://docs.openclaw.ai/help/faq/what-is-openclaw) (read 2026-09-30); [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins) (read 2026-09-30); [OpenClaw custom and local providers](https://docs.openclaw.ai/concepts/model-providers/custom-providers) (read 2026-09-30); [OpenClaw usage and costs](https://docs.openclaw.ai/reference/api-usage-costs) (read 2026-09-30); [OpenClaw tool and agent permissions](https://docs.openclaw.ai/gateway/security/tool-permissions) (read 2026-09-30) | ## What each product lets you control ### Claude, formerly Cowork Claude is merging Cowork and chat in stages. Its model picker remains, and cloud tasks can continue while your laptop is closed. Work with local files or apps needs Claude Desktop open. General third-party model keys are not established for this personal task experience. [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude); [Claude Cowork product](https://claude.com/product/cowork). Checked September 30, 2026. ### Dots Dots is OpenAI’s personal agent in ChatGPT, powered by GPT-6 Astra. Each dot has a cloud computer, with optional local access. The first dot comes with eligible Pro or Business Premium plans, but deeper work has an allowance. A Dots model picker and BYOK remain undocumented. [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/); [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot). Checked September 30, 2026. [Dots plans, setup, and permissions](https://benchlm.ai/blog/posts/openai-dots-guide) ### Grok Bot Grok Bot is the xAI product operated on Cursor-hosted computers. Cursor chooses the serving model; the product has no customer model picker. All Bots on one user account share files, browser sessions, and logins. Linked usage grants do not stack; the FAQ says the larger subscription allowance applies. Actual account charges have not been tested here. [Grok Bot hosting and model controls](https://docs.x.ai/grok-bot/security); [Grok Bot computer and access FAQ](https://docs.x.ai/grok-bot/faq); [Grok Bot plans and billing](https://cursor.com/help/grok-bot/plans). Checked September 30, 2026. ### Hermes Agent Hermes Agent is MIT-licensed software you can run yourself. Model selection and provider credentials are configurable, including compatible local endpoints. Nous Portal credits and Hermes Cloud hosting are optional services with separate meters. Native Anthropic OAuth requires Max plus extra usage credits; it does not draw on the base Max allowance. [Hermes Agent licensing and hosting](https://hermes-agent.nousresearch.com/); [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers); [Hermes Cloud billing](https://portal.nousresearch.com/cloud). Checked September 30, 2026. ### Manus Manus 2.0 runs agent work on managed infrastructure and offers an authorized local computer-use session. Its Cloud Computer is a paid add-on for projects needing a persistent environment. Current model selection, BYOK, and that add-on’s price are unverified; older Manus 1.6 labels do not establish 2.0 restrictions. [Manus 2.0 and computer use](https://manus.im/blog/introducing-manus-2-0); [Manus credit plans](https://manus.im/pricing); [Manus task-credit billing](https://help.manus.im/en/articles/11711097-what-are-the-rules-for-credits-consumption-and-how-can-i-obtain-them). Checked September 30, 2026. ### OpenClaw OpenClaw’s Gateway runs on your own infrastructure. You choose provider authentication and models, including documented local-runtime options. A remote model or connected service can still receive task data. The software does not establish your operating bill: inference, optional tools, and the machine remain separate costs. [OpenClaw runtime and trust boundary](https://docs.openclaw.ai/help/faq/what-is-openclaw); [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [OpenClaw custom and local providers](https://docs.openclaw.ai/concepts/model-providers/custom-providers); [OpenClaw usage and costs](https://docs.openclaw.ai/reference/api-usage-costs). Checked September 30, 2026. ## Choose by the requirement that cannot change Choose the requirement that would rule a product out before comparing plans. If you must select the model provider or connect a local endpoint, start with OpenClaw or self-hosted Hermes. Budget for the host and check the chosen authentication route; neither installation makes remote inference local. [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [OpenClaw custom and local providers](https://docs.openclaw.ai/concepts/model-providers/custom-providers); [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers). For a managed product, check your existing account first: can it access the runtime, connect the required sources, and permit the task? Compare Dots, Grok Bot, Claude, and Manus against those requirements in the table. Reject any option whose required control remains unknown until its documentation or your account confirms it. Then try one task you can inspect before paying for more usage. This is our selection approach, not a tested performance ranking. ## Read the bill behind the plan A subscription buys an allowance under that provider’s rules. A provider API key can add an inference bill. Search, browser, and other connected tools can add charges, and a machine running the agent adds hosting or hardware costs. Hermes Cloud explicitly bills inference and tools on top of compute and storage. [Hermes Cloud billing](https://portal.nousresearch.com/cloud). Keep those amounts in separate lines of your budget. Different providers’ credits do not measure the same work, and a stopped agent can still need paid storage. An unknown charge is not zero. No complete monthly or per-task bill has been measured here. For the model portion, use the [LLM API price table](https://benchlm.ai/llm-pricing) and [model token-cost calculator](https://benchlm.ai/tools/cost-calculator). They do not estimate the whole personal-agent bill. ## Check access and permissions before choosing First confirm the product is available to your account and region. Dots is rolling out gradually, and Claude accounts can differ during its transition. A subscription name alone does not guarantee the new task experience. [Dots access and controls](https://help.openai.com/en/articles/20001530-getting-started-with-your-dot); [Claude transition and task controls](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude). Then check the boundary you are granting: selected folders, connected-app permissions, a logged-in cloud browser, or local command execution. Grok Bots on one account share a computer and its logins. Creating another Bot does not isolate those credentials. [Grok Bot computer and access FAQ](https://docs.x.ai/grok-bot/faq). Self-hosting tells you where the runtime lives. It does not tell you where every request goes: remote inference and connected services can receive data. Check the chosen model route and tools before treating a local installation as a local-only setup. [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers). ## Guides and tests available for each product The [Dots guide](https://benchlm.ai/blog/posts/openai-dots-guide) explains its documented plans and controls. The [OpenClaw, Hermes, and Grok Bot overview](https://benchlm.ai/blog/posts/we-measure-the-session-they-sell-the-week) discusses persistent work. Neither is a matched product test; dedicated setup and billing guides for the self-hosted products are planned. For model selection, the [agentic model shortlist](https://benchlm.ai/best/agents) and [agent benchmark results](https://benchlm.ai/llm-agent-benchmarks) cover model and harness evidence. Those scores do not rank the six personal-agent products above. ## Sources, corrections, and freshness We include identifiable personal-agent products with primary documentation about how they run. Products are listed alphabetically. Each cell links to the source checked for that claim, including sources inspected for an unanswered question. “Unknown” means the checked material does not settle the field; it does not mean the feature is absent. “Not tested” describes our product-test coverage. Current setup and billing documentation takes precedence over broad launch descriptions. Documented allowances stay separate from an untested account outcome. Prices identify the currency, unit, and billing interval; unverified amounts stay out of the table. Read dates describe source checks, not successful onboarding or invoices. Billing, eligibility, and rollout can change after September 30, 2026. Send a source URL and the affected cell through [the correction form](https://benchlm.ai/contact) so we can check it. ## Questions ### Can a self-hosted agent use a remote model? Yes. OpenClaw and Hermes document remote provider routes alongside local options. Hosting the runtime yourself does not keep remote inference or connected-app requests on your machine. Check the exact endpoint, authentication route, and tool settings before deciding what data leaves the host. [OpenClaw provider authentication](https://docs.openclaw.ai/concepts/model-providers/official-provider-plugins); [Hermes model providers and authentication](https://hermes-agent.nousresearch.com/docs/integrations/providers). ### Does a subscription include every agent charge? It depends on the route and allowance. Dots delegates Codex and ChatGPT Work tasks to their usual usage meters. Hermes Cloud adds hosting and storage charges separately from inference and tools. Extra usage and connected services can change the bill, so an included agent is not a measured all-in operating cost. [OpenAI Dots launch and usage](https://openai.com/index/introducing-dots/); [Hermes Cloud billing](https://portal.nousresearch.com/cloud). ### Which personal agent performs best? This comparison has no matched six-product task test, so it cannot establish a performance winner. The table can help you choose a configuration by hosting, model control, billing, and permissions. [Model and harness benchmarks](https://benchlm.ai/llm-agent-benchmarks) are separate evidence; they do not measure every product’s persistent work or complete job cost.