Issues

• [Bug]: is_gpt_reasoning_series_name checks raw model path instead of normalized name, misclassifying models under gpt-5/gpt-6 routing prefixes

  llm translation
  BerriAI/litellm#45998 · Filter by author chauvuusvn opened on Oct 12, 2026

• [Feature]: Continue Responses fallback when partial usage extraction fails

  enhancement
  BerriAI/litellm#45993 · Filter by author LiuQiang-stack opened on Oct 12, 2026

• [Bug]: Gemini tool-use prompt tokens are added to prompt_tokens but not billed

  llm translation
  BerriAI/litellm#45991 · Filter by author porkedchop opened on Oct 12, 2026
  2

• [Bug]: Startup schema check suggests an invalid primary-key rewrite for partitioned SpendLogs

  BerriAI/litellm#45988 · Filter by author emerzon opened on Oct 12, 2026

• [Bug]: router_settings.enable_pre_call_checks: true does not enforce deployment rpm on /v1/messages

  llm translation
  BerriAI/litellm#45987 · Filter by author xinliu0727-sudo opened on Oct 12, 2026
  2

• [Bug]: rerank on the huggingface provider (TEI) always costs $0, configured input_cost_per_query is ignored

  llm translation
  BerriAI/litellm#45975 · Filter by author Mak5ens opened on Oct 11, 2026
  1

• [Bug]: MCP REST OpenAPI loader rejects YAML documents

  BerriAI/litellm#45964 · Filter by author teenaxta opened on Oct 11, 2026

• [Bug]: RouterRateLimitError in litellm/router.py

  bugllm translation
  BerriAI/litellm#45949 · Filter by author 4-FLOSS-Free-Libre-Open-Source-Software opened on Oct 11, 2026

• [Bug]: Per-request fallbacks in a multipart /v1/images/edits request is iterated one character at a time

  bug
  BerriAI/litellm#45948 · Filter by author richjun opened on Oct 11, 2026
  1

• [Bug]: A2AConfig declares temperature, max_tokens and top_p as supported then silently drops all three

  llm translation
  BerriAI/litellm#45944 · Filter by author JingHao-Leon opened on Oct 11, 2026
  5

• [Bug]: With custom_auth_run_common_checks: true and no database, key max_budget pre-call reservation is never reconciled on success

  llm translation
  BerriAI/litellm#45895 · Filter by author xinliu0727-sudo opened on Oct 11, 2026
  1

• [Bug]: per_user_oauth Copilot models in responses mode fail with "requires litellm_credential_name"

  claude codellm translation
  BerriAI/litellm#45883 · Filter by author leonnardo opened on Oct 10, 2026

• Opt-in request security middleware for the proxy (IP blocklists, general rate limiting, payload screening)

  BerriAI/litellm#45872 · Filter by author rennf93 opened on Oct 10, 2026

• [Feature]: Let a session-affinity pin yield when its deployment is saturated and another deployment in the group is idle

  BerriAI/litellm#45868 · Filter by author lobo235 opened on Oct 10, 2026
  3

• [Bug]: Anthropic batch accounting undercounts 1h cache writes and omits web-search fees

  llm translation
  BerriAI/litellm#45830 · Filter by author mavericksea-ai opened on Oct 10, 2026

• [Bug]: Getting 500 error

  bug
  BerriAI/litellm#45825 · Filter by author adarsha193 opened on Oct 10, 2026

• [Feature]: Model Additon: Microsoft Decison 1

  enhancementllm translation
  BerriAI/litellm#45807 · Filter by author wagnerjt opened on Oct 10, 2026
  1

• [Bug]: OpenTelemetry callback records tool-call responses as an empty message (gen_ai.output.messages has "parts": []), and "stop" as the finish reason for single-chunk streamed tool calls

  bugllm translation
  BerriAI/litellm#45796 · Filter by author armanunix opened on Oct 10, 2026
  2

• [Bug]: Pass-through endpoint URL containing 'predict' substring is misidentified as Vertex AI route — KeyError('predictions') kills logging task, SpendLogs never written

  llm translation
  BerriAI/litellm#45787 · Filter by author yelnyboy opened on Oct 10, 2026
  1

• [Bug]: Vertex AI serves a STALE access token up to the second it expires, Google answers 401 ACCESS_TOKEN_EXPIRED

  llm translation
  BerriAI/litellm#45785 · Filter by author mindw opened on Oct 10, 2026

• [Bug]: OpenTelemetry callback creates no span for a streaming call the app stops reading early

  bugllm translation
  BerriAI/litellm#45736 · Filter by author armanunix opened on Oct 10, 2026
  4

• [Bug]: routing_strategy "cost-based-routing" breaks all sync Router methods with RouterRateLimitError (single healthy deployment) — permanently disables mcp_semantic_tool_filter

  llm translation
  BerriAI/litellm#45718 · Filter by author automatCZ opened on Oct 10, 2026
  3

• [Bug]: /v1/messages does not accumulate deployment-level tpm for native anthropic/... deployments

  llm translation
  BerriAI/litellm#45702 · Filter by author xinliu0727-sudo opened on Oct 10, 2026
  2

• [Bug]: Counting tokens for a base64 PDF has no cost bound, so a few uploads stall every token count on the worker

  bugllm translation
  BerriAI/litellm#45676 · Filter by author RachelHuangZW opened on Oct 9, 2026
  2

• [Bug]: Claude models on Vertex AI fail at batch creation (404 batchPredictionJobs) — OpenAI batch translation doesn't cover Anthropic's native schema

  llm translation
  BerriAI/litellm#45671 · Filter by author manarayanareddy-clgx opened on Oct 9, 2026
  2