GPT-5 nano (and other GPT-5 models) use reasoning that consumes
the output token budget. When max_tokens is too low, all tokens
get used by internal reasoning, leaving nothing for the response.
- Add needsHigherTokenLimit() to detect models needing more tokens
- Add getMinTokenLimit() to ensure minimum 16k tokens for GPT-5
- Update buildCompletionParams to apply minimum token limits
- This fixes the No response from AI error with gpt-5-nano
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add AIUsageLog table migration for token tracking
- Fix GPT-5 temperature parameter (not supported, like o-series)
- Add usesNewTokenParam() and supportsTemperature() functions
- Add GPT-5+ category to model selection UI
- Update model sorting to show GPT-5+ first
GPT-5 and newer models use max_completion_tokens and don't support
custom temperature values, similar to reasoning models.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
GPT-5 and newer models require max_completion_tokens instead of max_tokens.
Added usesNewTokenParam() to detect GPT-5+ models separately from reasoning
model restrictions (temperature, json_object, system messages).
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add listAvailableModels() and validateModel() to openai.ts
- Improve testOpenAIConnection() to test configured model
- Add checkAIStatus endpoint to filtering router
- Add pre-execution AI config check in executeRules
- Improve error messages in AI filtering service (rate limit, quota, etc.)
- Add AI status warning banner on round detail page for filtering rounds
Now admins get clear errors when AI is misconfigured instead of silent flags.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add getConfiguredModel() that reads ai_model from SystemSettings
- AI assignment and mentor matching now use the admin-selected model
- Remove duplicate OpenAI client in mentor-matching (use shared singleton)
- Add GPT-5, GPT-5 Mini, o3, o3 Mini, o4 Mini to model dropdown
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Full Next.js 15 platform with tRPC, Prisma, PostgreSQL, NextAuth.
Includes production Dockerfile (multi-stage, port 7600), docker-compose
with registry-based image pull, Gitea Actions CI workflow, nginx config
for portal.monaco-opc.com, deployment scripts, and DEPLOYMENT.md guide.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
