feat: add Telnyx provider

[regismesquita]

Summary

• add Telnyx as an OpenAI-compatible inference provider
• add Kimi K2.6, Kimi K3, and GLM-5.2 with provider-specific pricing, limits, modalities, and reasoning controls
• add a square currentColor Telnyx logo based on the official GitHub organization avatar

Sources

• Models: https://developers.telnyx.com/docs/inference/models
• Pricing: https://telnyx.com/pricing/inference-api
• Kimi K3 reasoning: https://telnyx.com/resources/kimi-k3-open-model-telnyx-inference
• Logo reference: https://avatars.githubusercontent.com/u/10522416?s=200&v=4

Verification

• bun validate
• live API Dev chat-completions requests succeeded for all three model IDs
• live API Dev tool-calling requests using tools together with max_completion_tokens returned one tool call for all three models
• OpenCode local catalog resolves GLM-5.2 reasoning variants as high and max

Scope

Only providers/telnyx/ is included. MiniMax-M3 is intentionally excluded because it requires a separate Anthropic Messages endpoint and protocol-specific configuration.