Grok Bot vs Claude Cowork: plans, prices and differences
Grok Bot starts at xAI's $30 Premium plan and Claude Cowork at Claude Pro's $20 a month. Both run work in the cloud, and neither company publishes usage numbers.
770 stories
Grok Bot starts at xAI's $30 Premium plan and Claude Cowork at Claude Pro's $20 a month. Both run work in the cloud, and neither company publishes usage numbers.
ServiceNow AI's open model hears audio, reasons and calls tools in one network. It beats some closed rivals and trails Gemini 3.8 on its own tests.
xAI's Grok Bot runs persistent AI teammates on a cloud computer. Here are the steps for the app, access through your plan, and tagging @bot on X.
Grok Bot comes with Premium, Super and Ultra, a linked SuperGrok subscription or a Grok Business seat. xAI states no separate price, and weekly limits are tightening.

Public repositories lose the trigger unless a policy allows it, with test windows on November 9 and 16 first.
A paper from Chinese University of Hong Kong, Shenzhen uses self-play to move hand-written skills into a Director model, and reports wins on 12 benchmarks.
Midjourney sells a subscription workspace with a version 8.2 model. Google's Nano Banana is billed per image in the API, from $0.039 to $0.24.
Batch your messages, clear context, trim CLAUDE.md and MCP servers, and match the model and effort to the task, according to Anthropic's docs.
Work is for finished files, Codex is for code. They run on the same engine, work in the same plans and share one usage allowance.
Select users get the combined tab first. Codex stays separate, and Chat and Codex share one usage pool.
The updated IPO filing adds third-quarter results and discloses $895.4 billion in purchase commitments, almost all for compute.
The fixed 4.5-for-1 share swap pairs Marvell's AI chips with Lumentum's lasers and optical parts, and is expected to close in the first half of 2028.
A Tsinghua perspective paper says a cheap predictor could decide what an expensive executor runs. Its savings figures are theoretical bounds, and it reports no experiments.
Gemini Notebook, formerly NotebookLM, has no plan of its own. Limits scale with Google AI Plus, Pro and Ultra, from 100 notebooks free to 500 notebooks with 600 sources.
OpenAI publishes the Codex CLI on GitHub under Apache 2.0, but it takes no outside code, and running it still needs a ChatGPT plan or an API key.
Check status.x.ai first, then match the error code to the fix in xAI's own docs, from a 401 key problem to a 429 rate limit.

Opus 5.5 costs 60% less and tops Fable 5.1 on every row of Anthropic's own benchmark table. Anthropic still keeps Fable for the hardest agent work.
Two routes from Anthropic's docs: the Agent SDK for a ready-made loop and built-in tools, or the Messages API when you want to write the loop yourself.
Google renamed NotebookLM to Gemini Notebook in July 2026. Here is how its sources, limits and plans compare with Claude.
OpenAI's status page showed no incident on November 2, 2026. Here is the order to try the fixes OpenAI's help center lists.
DeepSeek's docs describe no free tier or credit amount. Usage is deducted from a topped-up or granted balance, and requests fail with a 402 when it hits zero.
SemiAnalysis tests frontier models on 206 expert tasks in 11 fields, and finance and legal work mostly goes unsolved.
Meta's help pages describe no off switch in WhatsApp, Instagram, Facebook or Messenger. They do cover deleting chats, WhatsApp incognito chats and a glasses setting.
Check the status page, hard refresh, drop the VPN, then confirm your account email. Perplexity's help center has a fix for each common failure.