chore(chutes): sync models with live API

[TheStreamCode]

Summary

The Chutes provider listing had drifted from the live llm.chutes.ai/v1/models API — newer models were missing and retired ones lingered. This syncs the provider to the API's current set (13 models).

Changes

Added (3)

Live on the API but missing from the catalog:

• zai-org/GLM-5.2-TEE
• Qwen/Qwen3-235B-A22B-Thinking-2507-TEE
• unsloth/Mistral-Nemo-Instruct-2407-TEE

Updated (8)

Refreshed pricing/limits from the API (TEE-specific manual fields such as reasoning_options and interleaved preserved):
Qwen3-32B-TEE, Qwen3.5-397B-A17B-TEE, Qwen3.6-27B-TEE, DeepSeek-V3.2-TEE, gemma-4-31B-turbo-TEE, Kimi-K2.5-TEE, Kimi-K2.6-TEE, GLM-5.1-TEE

Removed (29)

No longer served by the API:
GLM-4.7-TEE, GLM-4.7-FP8, GLM-4.6V, GLM-5-Turbo, Qwen2.5-72B-Instruct, Qwen2.5-Coder-32B-Instruct, Qwen2.5-VL-32B-Instruct, Qwen3-235B-A22B-Instruct-2507-TEE, Qwen3-235B-A22B-Thinking-2507, Qwen3-30B-A3B, Qwen3-Coder-Next-TEE, Qwen3-Next-80B-A3B-Instruct, Qwen3Guard-Gen-0.6B, MiMo-V2-Flash-TEE, DeepSeek-R1-0528-TEE, DeepSeek-R1-Distill-Llama-70B, DeepSeek-V3-0324-TEE, DeepSeek-V3.1-TEE, gpt-oss-120b-TEE, dots.ocr, DeepSeek-TNG-R1T2-Chimera-TEE, Llama-3.2-1B-Instruct, Llama-3.2-3B-Instruct, Mistral-Nemo-Instruct-2407, gemma-3-12b-it, gemma-3-27b-it, gemma-3-4b-it, DeepHermes-3-Mistral-24B-Preview, Hermes-4-14B

After this change the provider lists exactly the 13 models the API returns.

Note: GitHub renders a few of the add/remove pairs as renames due to content similarity (e.g. GLM-4.7-TEE.toml → GLM-5.2-TEE.toml); these are independent add/remove operations.

Validation

bun run validate passes.

[rekram1-node]

can we easily add a sync script?

[TheStreamCode]

Yes — it'd just be a standard src/sync/providers/chutes.ts module (today generate-chutes.ts is one of the standalone generators not wired into bun models:sync, which is why this drifted). /v1/models is public, so no new secret would be needed.