Problem (one or two sentences)

When use API Provider Anthropic and choose custom url, i want to make a custom model too and not to be block to only the current list of model who are only claude-xxx models

Because i can use an endpoint like https://api.z.ai/api/anthropic with model like glm-4.6-cc-max or glm-4.5v who are really more speeder like https://api.z.ai/api/coding/paas/v4 (who are OpenAi compatible endpoint)

With that i can make too a profile to make a quick change to vision model capabilities instead of using mcp server for it

ruby .\llm_benchmark.rb --provider z-ai --model glm-new
=== LLM Benchmark ===
Provider: z-ai (https://api.z.ai/api/coding/paas/v4)
Model: glm-new (glm-4.6)
Starting benchmark...
Start time: 2025-10-03 14:27:56.436
End time: 2025-10-03 14:28:08.199

=== Results ===
Duration: 11.763 seconds
Input tokens: 45
Output tokens: 1000
Total tokens: 1045
Tokens per second: 88.84
PS C:\Users\cobra.claude> ruby .\llm_benchmark.rb --provider z-ai-anthropic --model glm-new
=== LLM Benchmark ===
Provider: z-ai-anthropic (https://api.z.ai/api/anthropic/v1)
Model: glm-new (glm-4.6)
Starting benchmark...
Start time: 2025-10-03 14:28:25.890
End time: 2025-10-03 14:28:34.620

=== Results ===
Duration: 8.729 seconds
Input tokens: 43
Output tokens: 1000
Total tokens: 1043
Tokens per second: 119.48

Context (who is affected and when)

user of z.ai coding plan

Desired behavior (conceptual, not technical)

able to write any model instead of only the initial list

Constraints / preferences (optional)

No response

Request checklist

• I've searched existing Issues and Discussions for duplicates
• This describes a specific problem with clear context and impact

Roo Code Task Links (optional)

No response

Acceptance criteria (optional)

No response

Proposed approach (optional)

No response

Trade-offs / risks (optional)

No response