[BUG] Something went wrong Try sending your message again. If it keeps happening, share feedback so we can investigate. The session stopped responding. Send your message again to resume with a fresh process.

Found 3 possible duplicate issues:

1. https://github.com/anthropics/claude-code/issues/39869
2. https://github.com/anthropics/claude-code/issues/41802
3. https://github.com/anthropics/claude-code/issues/52860

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

Each attempt consumes a lot of tokens, but it fails. WTF?

I continued working in the CLI. Unlike the desktop version, it was able to complete the entire request.

Sent a feedback from CLI so the team could analyze the cause
Feedback ID: 7b48ec22-51cf-41ac-ba79-8c44d4d40ab0

Same issue. It has just become completely unusable to work. Literally, no agent is successfully completing its task, and this (the same error in your screenshot) is just being thrown away at the screen every time.

Having the same exact issue, I tried rebooting my app but it seems to be a model related issue. I think the only way to fix is to start a new chat?

Same issue in multiple devices last few days

Same issue here. Happening A LOT lately

Same issue, started today

Same issue, started today
Version 1.7196.0 (2dbd78)

same issue

claude for windows 1.7196.0 (2db78)
opus4.7 (medium)

98549bbe-76ce-45d5-8ad5-4b1a2e25c101

same issue

mac issue the same

same issue - cowork is now completely unusable, and I can't tell you what a real bummer that is bc of all of the work I had tied up in it. It was magical until suddenly it wasn't. This started for me middle of the week, ironically while I was at a Claude for SMB event. Just suddenly with every task chat in CoWork. I also in the last two weeks ran into the usage limits early for both weeks on my $200 Claude Max plan (and I'm seriously considering cancelling it when I'm done with a bunch of builds in Claude Code). And no, I wasnt using sdk or trying to use OpenClaw or anything like that. Just building stuff and working. As you can imagine, this is a bit of a nightmare. I had grown ever more reliant on Claude over the last three months for my business, and thought it was going to my forever app for at least the next few years. But not anymore. As a result, I have quickly pivoted to an explosion of other work this weekend. Friday bc I was locked out bc of usage limits, and bc Cowork was suddenlty useless, I started using Codex for the first time, a fortuitous turn as it has grown truly robust. And I also started using Antigravity for the first time. Are they as intuitive or as good as Claude Code? No. Are the results as good? No, but they didn't lock me out either, and the results are not bad. In the last three days I have also installed paperclip.ing and am using it via other models as a multi-agentic platform for my business. On top of this, I've set up a Kimi subscription as of today, installed an OpenClaw for the first time ever via Kimi Claw (their new feature), and this morning I installed Hermes Agent and have been using Kimi's API to work with Hermes via Terminal. I was using Claude Sonnet for a little bit, and guess what? Just chewed through API. But it barely sips Kimi. I also installed Open Design this weekend and used it. Worked incredibly well. Lastly I began using Perplexity Personal Computer. It's a token hog also but the way it swarms with tasks and agents is truly incredible. It's been an eye opening weekend.

我也频繁遇到这个问题. 补充一些细节:

环境:

• 平台: macOS (Darwin 25.5.0)
• Shell: fish
• 模型: deepseek-v4-pro:cloud 和 GLM-5.1 都出现过
• 版本: Claude Code CLI

症状:

• 对话进行到中途, Claude 输出被截断, 停在 "Thought for Xs" 或 "Cogitated for Xs" 等进度信息
• 没有任何错误信息, 没有超时报错, 就是静默停止
• 多次出现, 每次都需要手动重新输入才能继续
• 不是 context 用完 (有时候 compact 完立刻又出现)

触发模式 (观察):

• 长对话 + 多次 tool call 之后更容易出现
• 上下文压缩后似乎更容易触发
• 并发工具调用时也会出现

影响:

• 打断开发流程, 非常影响体验
• 用户需要手动多次重新输入才能继续工作

希望能帮助定位.

Try sending your message again. If it keeps happening, share feedback so we can investigate.

API Error: Usage credits required for 1M context · turn on usage credits at claude.ai/settings/usage, or use --model to switch to standard context
You can restart the conversation from an earlier message.

У меня такая же проблема,ничего не помогает

Closing for now — inactive for too long. Please open a new issue if this is still relevant.