feat(vancine): refresh DeepSeek models and GLM pricing

[fx247562340]

Summary

• Replace the retired Vancine deepseek-v4-flash, deepseek-v4-pro, and deepseek-v4-flash-vision-exp catalog entries with deepseek-flash (DeepSeek V4.1 Flash).
• Publish Vancine pricing for deepseek-flash: $0.24 input, $0.96 output, and $0.0048 cache read per million tokens.
• Update Vancine GLM-5.3-Flash pricing after the half-price promotion ended: $0.12 input, $0.40 output, and $0.024 cache read per million tokens.
• Preserve the model’s inherited capabilities and Vancine-specific reasoning controls.

Sources

• Vancine public pricing: https://vancine.com/api/pricing
• Vancine Pi catalog metadata: https://vancine.com/api/pi/catalog

Prices were verified on 2026-09-11.

Validation

• bun validate

AI assistance disclosure

This change was prepared with AI assistance and reviewed and validated by the contributor.

[github-actions]

No actionable findings.