feat: add new Chutes TEE providers and fix model configuration fields

[mark182es]

Summary

This PR adds support for new TEE (Trusted Execution Environment) providers and fixes missing configuration fields in chutes models.

New TEE Providers

• MiniMaxAI: MiniMax M2.1 TEE
• NousResearch: Hermes 4 405B FP8 TEE
• Qwen: Qwen2.5 VL 72B Instruct TEE, Qwen3 235B A22B Instruct 2507 TEE, Qwen3 Coder 480B A35B Instruct FP8 TEE, Qwen3 Guard Gen 0.6B
• deepseek-ai: DeepSeek R1 0528 TEE, DeepSeek R1 TEE, DeepSeek V3 0324 TEE, DeepSeek V3.1 TEE, DeepSeek V3.1 Terminus TEE, DeepSeek V3.2 TEE
• moonshotai: Kimi K2 Thinking TEE
• openai: GPT OSS 120B TEE
• zai-org: GLM 4.5 TEE, GLM 4.6 TEE, GLM 4.7 TEE
• XiaomiMiMo: MiMo V2 Flash
• NVIDIA: Nemotron 3 Nano 30B A3B BF16

Fixes

• Add missing structured_output field to all chute models
• Fixed context length for all models
• Fixed pricing for all models

Coverage

• 62/62 models updated (100% coverage)
• 55 models with structured_output = true
• 35+ models with reasoning = true (including interleaved section)
• All models maintain API accuracy

Source of data

• https://llm.chutes.ai/v1/models

Breaking Changes

⚠️ None - All changes are additive and maintain backward compatibility:

• Existing model IDs remain unchanged
• TEE models coexist with regular versions
• No changes to provider API or authentication

Future improvements

• Auto update with the Chutes API for new models added, pricing and remove inactive models

Testing

• All TOML files validated for correct syntax
• All required schema fields are present
• Provider and model naming conventions followed

---

First contribution to open source! 🚀

This is my first PR to this project. Feedback and suggestions are welcome!