Reports show some heavy users moving to SpaceXAI’s $300 SuperGrok Heavy tier and adopting Grok Bot agents after slowdowns in competing models and because Grok’s agent features and pricing match agentic workflows.
Grok 4.6 is cited for balancing intelligence with lower per-task costs on agentic, multi-step tasks.
Grok Bot provides always-on, persistent agents that accumulate memory, connect to many integrations, and can be taught by demonstration.
xAI co-founder Jimmy Ba says most xAI engineers use agent loops and run 10–20 GrokBot agents to automate routines and speed development.
# Why some heavy users are switching
# What Grok Bot changes about agent workflows
Grok Bot is described as an always-on agent system that represents agents like teammates in an iMessage-style interface. A bot is treated as a persistent role: it keeps memory, owns a domain, and improves through repeated interactions in the same thread. That persistence matters for routines you run frequently.
Grok Bot's interface includes a plugin panel for integrations. Available connectors mentioned include Notion, Slack, Google Drive, AWS Agents, AWS SageMaker, Browserbase, Composio, and Context7, plus an option to build a custom integration. Connecting an integration once makes it available to all bots.
You can teach a Grok bot by demonstration: perform a workflow while the bot watches, and the bot saves the routine. It can then execute the same steps on its own. The recommended candidate tasks are those you do at least weekly, use two or more tools, and have stable steps. Examples given are scheduling, email management, daily briefings, weekly summaries, and month-end expense filing.
# How engineers are using Grok agents internally
xAI co-founder Jimmy Ba is quoted describing internal usage: he says 90% of xAI engineers use AI agent loops to speed shipping, with each engineer running 10–20 GrokBot agents to automate routine work. That claim frames Grok Bot not as a one-off assistant but as a software development multiplier inside xAI's workflows.
# Material tradeoffs that users mention
Users praising Grok 4.6 point to two concrete advantages: (1) lower per-task cost on agentic tasks compared with alternatives, and (2) a sweet spot of intelligence for common multi-step work. Some users still prefer Claude for coding tasks, so the shift is not uniform across all workloads. Cancelations of Claude subscriptions by some heavy users are reported, signaling a partial market movement rather than a complete migration.
# Practical guidance for readers evaluating a switch
# Bottom line
SpaceXAI has released Grok 4.7, its new flagship model for coding, agentic tasks, and knowledge work. Grok 4.7 is built on a larger base model and a longer reinforcement learning run. It still ships at the same price and speed as Grok 4.6. Is it deployable? Yes, as a hosted model. You can call grok-4.7 […] The post Spa
Grok 4.7 promises stronger AI coding performance without a higher base price, but heavy token consumption could reduce those savings. SpaceXAI released Grok 4.7 on Sept. 21 as its latest model for coding and professional knowledge work. According to the company, it uses a larger base model than Grok 4.6 and underwent a
SpaceXAIは21日(米国時間)、最新のAIモデルとなる「Grok 4.7」を公開した。コーディングやナレッジワーク向けで、同社における最高モデルとなり、同等のモデルと比較し、「処理速度は2倍、価格は半額」とアピールしている。CursorとGrok Buildで利用可能となるほか、Grok APIや各社のコーディングハーネス、モデルルーター、クラウドプラットフォームでも展開する。
Last month, SpaceXAI launched Grok Bot, an iMessage-style AI system that displays agents as teammates. The product originally arrived as an iPhone and Mac app with a very expensive price tag. In the weeks since launch, Grok Bot has expanded to the iPad and dramatically lowered the cost. more…
SpaceXAI Grok 4.7 is here. It is way better than Grok 4.6 at the same price and speed. It is on the pareto frontier. The new model builds on a larger base with extended training on complex tasks, boosting scores like 46.3% on CursorBench 4.0 for software engineering and 71.0% on DeepSWE v1.1. It keeps ... Read more
Loading more related stories...
Open the app view to save this story, compare related coverage, and continue from the same source.