Since Oct 21 many developers on X have complained that Claude Code was draining usage limits much faster. On Oct 28 (UTC) Boris Cherny, who leads Claude Code at Anthropic, replied on X that the team reviewed every session ID users shared, found no single root cause and no regression in Claude Code, and is rolling out fixes for a few smaller issues that made some sessions burn tokens faster. Most fast-drain cases traced back to one or two heavy sessions using xhigh effort, /goal, Workflows, Agent Teams or long SDK jobs.
Key Takeaways
- ✓The complaint post that started it said usage was draining 5x faster and drew about 1,660 likes and 240+ replies (a user claim, not official data)
- ✓Boris Cherny: the team checked every shared session ID, found no single root cause or Claude Code regression, and is rolling out fixes for a few smaller issues
- ✓Heavy features were the main driver: xhigh effort, /goal, Workflows, Agent Teams and long SDK jobs
- ✓Subagents inherit the session effort level by default, which adds up fast; /usage shows where tokens went
- ✓The official docs say Agent Teams use about 7x more tokens than standard sessions when teammates run in plan mode
Key Decision Metrics at a Glance
Heavy Claude Code use: compare subscription limits and API bills
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Background and pain point
On Oct 21 developer Kaleb Campbell posted on X that Claude Code usage was draining 5x faster than before. The post drew about 1,660 likes and 240+ replies from subscribers reporting the same thing. For developers on plan limits, running out mid-task breaks their work. On Oct 22 Claude Code lead Boris Cherny replied that the team was investigating and asked users for session IDs.
What Anthropic found
On Oct 28 (UTC) Cherny posted the result in the same thread. The team reviewed every session ID shared and found a few smaller issues that made some sessions burn tokens faster than they should; fixes are rolling out now. It found no single root cause and no regression in Claude Code itself. Most fast-drain cases traced back to one or two heavy sessions using xhigh effort, /goal, Workflows, Agent Teams or long SDK jobs. He called these legitimate uses and said it is on Anthropic to make clearer when to use them and what they cost. Note that this is an employee's own post on X, not a formal Anthropic announcement.
Benchmarks and numbers
Anthropic has not published how many sessions were affected, how many extra tokens the bugs cost, or before-and-after data. The 5x figure is the original poster's claim and is unverified. The one documented reference point: Anthropic's docs say Agent Teams use about 7x more tokens than standard sessions when teammates run in plan mode, because each teammate keeps its own context window.
What developers can do now
Run /usage to see where a session's tokens went; it attributes usage to skills, subagents, plugins and MCP servers, and when it shows a hint you can press f to apply its recommendations. Watch your effort level: subagents inherit it by default, so many subagents at xhigh add up quickly, and /effort lowers it for simpler work. Agent Teams are off by default, so estimate costs before enabling them. If usage still looks wrong, send /feedback; Cherny says the team will follow up directly with everyone who shared a session ID.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.