Pi on an Anthropic subscription login cannot run Claude Opus 5.5 (claude_code_version_too_old)
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Newbie friendliness
- 73/100
- Issue type
- Bug
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- typescript
- Domain
- backend
Research direction
In services/runner, inspect the Pi dependency versions and services/runner/src/engines/sandbox_agent/pi-model-config.ts; the issue identifies the outdated OAuth version in pi-ai’s dist/api/anthropic-messages.js. Find a release presenting Claude Code 2.1.280 or newer, then update the dependencies and run the sync-model-catalog skill. Done when the subscription-login Opus 5.5 request works and the model catalog reflects the released model.
Written by the indexing model from the issue text.
Description
🤖 The AI agent says:
Reproduction
- Run a self-hosted Agenta with the Pi harness (
pi_core) authenticated through an Anthropic subscription login (Claude OAuth), not a vault API key. - Create an agent on the Pi harness with the model
anthropic/claude-opus-5-5. - Send any message.
The same failure reproduces outside the product by calling pi-ai 0.85.1's Anthropic streamSimple directly with a subscription token and the Opus 5.5 model.
Expected
The turn runs on Claude Opus 5.5, as it does with a vault Anthropic API key.
Actual
Anthropic rejects the request:
400 invalid_request_error: Claude Code 2.1.251 does not support this model; version 2.1.280 or newer is required.
error_code: claude_code_version_too_old
Cause
The runner pins @earendil-works/pi-ai 0.85.1 (via pi-coding-agent 0.85.1). On OAuth requests, pi-ai identifies itself as Claude Code 2.1.251 (const claudeCodeVersion = "2.1.251" in dist/api/anthropic-messages.js). Anthropic requires 2.1.280 or newer for Opus 5.5 on a subscription login.
Who is affected
- Affected: self-hosters who run the Pi harness on an Anthropic subscription login and pick Opus 5.5. The other models, including Fable 5.1, were not tested on that login, so they may hit the same version gate.
- Not affected: Pi runs with a vault Anthropic API key. Opus 5.5 is served there, and since #7072 it is priced and runs with reasoning.
- Not affected: the Claude harness. It pins Claude Code 2.1.280 through
claude-agent-acp0.81.0.
Fix
Bump pi-ai / pi-coding-agent in services/runner to a release that presents Claude Code 2.1.280 or newer on OAuth. After the bump:
- Delete the Opus 5.5 entry in
PI_MODELS_AHEAD_OF_CATALOG(services/runner/src/engines/sandbox_agent/pi-model-config.ts) once the catalog carries the model. - Run the
sync-model-catalogskill to regenerate the Pi catalog.
This was deliberately not bumped for v0.120.0.
- Dominant language
- TypeScript
- Stars
- 4.8k
- Forks
- 690
- Avg merge
- 19h 50m
- Merged PRs (30d)
- 419
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from Agenta-AI/agenta
-
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
Agenta-AI/agenta#7468 · 1 comment · 1 reaction ·
Maintainers usually reply within 1 day
-
(bug) The Stop button disappears for a moment before an approval card showsPossibly taken @rishu685 claimed this 5 days ago. Openbug
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
Agenta-AI/agenta#7446 · 2 comments · 1 reaction ·
Maintainers usually reply within 1 day
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 83/100
Agenta-AI/agenta#7441 · 1 comment · 1 reaction ·
Maintainers usually reply within 1 day
-
bug frontend
Difficulty 2/5 1-3 hours Newbie friendliness 65/100
Agenta-AI/agenta#7426 · 1 comment · 1 reaction ·
Maintainers usually reply within 1 day
-
No CI job runs the mobile app's unit testsPossibly taken @Mahaboob-Ashraf claimed this 5 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 82/100
Agenta-AI/agenta#7326 · 2 comments · 1 reaction ·
Maintainers usually reply within 1 day
All issues in Agenta-AI/agenta
Similar issues
-
Difficulty 1/5 1-3 hours Newbie friendliness 70/100
siyuan-note/siyuan#20517 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 70/100
NaturalIntelligence/fast-xml-parser#890 · 1 comment ·
Maintainers usually reply within 3 days
-
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
zsviczian/obsidian-excalidraw-plugin#3000 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 68/100
pnpm/get.pnpm.io#68 ·
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
nuxt-modules/better-auth#578 · 2 comments ·
Maintainers usually reply within 1 day