- DeepSeek V3.1: reasoning false (not a reasoning model)
- GPT OSS 120B: reasoning true
- Remove deprecated GLM 4.6 model
- Update hints to 'Private and verifiable inference'
- Fix alias to match default model (GLM-4.7)
- Add near-ai-api-key to preferred-provider mapping
- Add NEAR AI to docs/providers/index.md and models.md
- Add note about model list changes + link to cloud.near.ai/models
- Add Near AI as OpenAI-compatible provider
- Models: DeepSeek V3.1, GPT-OSS-120B, Qwen3-30B, GLM-4.6, GLM-4.7
- Auth via API key (NEAR_AI_API_KEY or clawdbot auth)
- Privacy-focused inference using Intel TDX/NVIDIA TEE
- Base URL: https://cloud-api.near.ai/v1