vava-nessa/free-coding-models v0.5.19

on GitHub

Added

• nvidia: Added nvidia/nemotron-3-ultra-550b-a55b (S+, 1M context)
• groq: Added 7 new specialized models:
  • whisper-large-v3, whisper-large-v3-turbo (audio transcription)
  • canopylabs/orpheus-arabic-saudi, canopylabs/orpheus-v1-english (Arabic/English speech)
  • meta-llama/llama-prompt-guard-2-22m, meta-llama/llama-prompt-guard-2-86m (security guardrails)
  • openai/gpt-oss-safeguard-20b (safety)

Fixed

• nvidia: Corrected context windows for three models:
  • deepseek-ai/deepseek-v4-pro: 128k → 1M
  • deepseek-ai/deepseek-v4-flash: 128k → 1M
  • mistralai/mistral-small-4-119b-2603: 128k → 256k
• nvidia: Removed deprecated model z-ai/glm5 (replaced by z-ai/glm-5.1)

Changed

• Model counts: nvidia stable at 27, groq increased from 8 to 15.
• Total catalog size: 161 models across 16 providers.

Notes

• This update reflects the latest provider catalogs as of June 10, 2026.
• All existing models were verified against official docs; only minor configuration drift was found.
• The audit_state.json was updated with fresh fingerprints.