Add support for Qwen3-235B-A22B-Thinking-2507 on Chutes AI provider

[sudofugo]

Is your feature request related to a problem?

Roo Code currently lacks support for Qwen3-235B-A22B-Thinking-2507 on the Chutes AI provider. This model delivers state-of-the-art reasoning, coding, and long-context capabilities (256K tokens) but is inaccessible to Roo Code users.

Describe the solution you'd like

Add support for Qwen3-235B-A22B-Thinking-2507 model via the Chutes AI provider in Roo Code. Key requirements:

• Enable model selection in Roo Code's UI
• Support 256K token context length
• Handle the model's unique "thinking mode" output (structured reasoning + final answer)
• Ensure compatibility with agentic workflows (tool usage)

Describe alternatives you've considered

• Using other models with smaller context windows (e.g., 128K) but this limits processing of large documents/codebases.
• Accessing the model through external tools, but this breaks Roo Code's integrated workflow.

Additional context

Model Highlights:

• SOTA Performance: Top-ranked open-source thinking model on reasoning (AIME25: 92.3%), coding (LiveCodeBench: 74.1%), and science benchmarks.
• 256K Context: Processes entire codebases/research papers in one pass.
• Agent-Optimized: Excels at tool usage (BFCL-v3: 71.9%) and complex task automation.
• Technical Specs: 235B parameters (22B activated), MoE architecture, thinking-mode only.

Why this matters:
This model enables Roo Code users to solve highly complex tasks (competitive programming, academic research) with unmatched accuracy and context depth - currently unavailable in the ecosystem.

Reference:
Qwen3-235B-A22B-Thinking-2507 on Hugging Face