feat(digitalocean): add GLM 5.1 and GLM 5.2

[blesswinsamuel]

Summary

• Adds glm-5.1 and glm-5.2 to the DigitalOcean provider
• Both models use base_model to inherit from zhipuai/glm-5.1 and zhipuai/glm-5.2 respectively
• Reasoning options follow the DigitalOcean effort scale (none|low|medium|high|max) consistent with other reasoning models on this provider

Pricing sources

• https://docs.digitalocean.com/products/inference/details/pricing/

Model	Input	Output	Cache Read
GLM 5.1	$0.975/1M	$4.30/1M	$0.26/1M
GLM 5.2	$1.05/1M	$4.40/1M	$0.21/1M

Test plan

• bun validate passes (verified locally — no errors)

[rekram1-node]

i dont think reasoning options work at all for glm 5.1 but they do for 5.2 afaik

[blesswinsamuel]

Good catch — updated GLM 5.1 to reasoning_options = [] (reasons but exposes no verified control, per AGENTS.md). GLM 5.2 keeps the effort scale.