AE-3ZDD3QWC
Gemini tool-call arguments returned through an LLM gateway escape Japanese text as \uXXXX sequences
MULTILINGUAL FAILUREseverity: LOWcause: LIKELYoutcome: RESOLVED VERIFIEDconfidence: LOW
When a Gemini model made a function call whose arguments contained Japanese text, the gateway's OpenAI-compatible response carried the arguments with every non-ASCII character escaped as a Unicode sequence, unlike OpenAI's behaviour. The cause was json.dumps without ensure_ascii=False when converting Gemini function-call args; a merged fix with a test added the flag, and the reporter confirmed it works.
- Framework / agent
- LiteLLM · tool-calling LLM application via a multi-provider gateway
- Remediation attempts
- VERIFIED SUCCESS
- Recurrence
- not documented
- Source languages
- en, ja
- Updated
- 2026-09-29
Sources
- GITHUB ISSUE [Bug]: Gemini function call arguments escape non-ASCII characters as Unicode sequences — github.com/BerriAI/litellm, retrieved 2026-09-29
- GITHUB PULL REQUEST fix(gemini): Preserve non-ASCII characters in function call arguments — github.com/BerriAI/litellm, retrieved 2026-09-29
Symptoms
- Function-call arguments contain \u escape sequences instead of the original Japanese characters
The full record — root-cause evidence, every remediation attempt with its status and verification, failed attempts, patch references, verbatim quotes and recurrence — is a paid lookup (0.018 USDC via x402). Agents:
GET /api/v1/cases/AE-3ZDD3QWC. PricingSimilarity to your system is not implied. A remediation that worked in the documented context may not work in yours.