fix: cline-pass: deepseek-v4.1-flash missing from provider data

[github-actions]

Summary

Adds the missing ClinePass model cline-pass/deepseek-v4.1-flash. It is already returned by the Cline recommended-models catalog under clinePass, while the other 14 ClinePass models were already catalogued. The lab entry deepseek/deepseek-v4.1-flash already exists, so this is an override-only provider file with base_model.

Changes

• Added providers/cline-pass/models/cline-pass/deepseek-v4.1-flash.toml
  • base_model = "deepseek/deepseek-v4.1-flash"
  • reasoning_options: effort none / low / medium / high / xhigh (same surface as existing ClinePass DeepSeek entries)
  • cost (USD/MTok, off-peak reference): input 0.15, output 0.6, cache_read 0.003

Evidence

• Availability / model ID: Cline recommended-models API lists clinePass entry id: "cline-pass/deepseek-v4.1-flash" with description “Smarter and more efficient, with 1M context window”.
• Lab identity / limits: existing models/deepseek/deepseek-v4.1-flash.toml (1M context, reasoning, vision, etc.).
• Pricing basis: DeepSeek Models & Pricing — DeepSeek-V4.1-Flash off-peak $0.15 / $0.6 / $0.003 per 1M tokens (cache hit). ClinePass docs still only tabulate deepseek-v4-flash and footnote DeepSeek API pricing for quota reference; no dedicated V4.1 Flash row yet.
• Reasoning options on this host: existing providers/cline-pass/models/cline-pass/deepseek-v4-flash.toml and deepseek-v4-pro.toml use the same effort set on the ClinePass OpenAI-compatible surface.

Validation

• Confirmed catalog presence via live fetch of the recommended-models endpoint.
• Cross-checked lab metadata and sibling ClinePass DeepSeek TOMLs.
• Did not run bun validate (no shell/Bash in this workflow).

Review notes

• ClinePass public docs model table does not yet list cline-pass/deepseek-v4.1-flash; availability is verified from the live catalog API.
• Used DeepSeek off-peak Flash rates for quota reference (same basis as first-party deepseek-flash). Did not change the existing deepseek-v4-flash ClinePass costs, which differ from both current DeepSeek off-peak and the ClinePass docs Flash rows.
• Matched ClinePass peer DeepSeek effort values rather than first-party DeepSeek toggle + low/high/max, for consistency on this host.

Closes #6861

Automated by the issue fixer: https://github.com/anomalyco/models.dev/actions/runs/34577702256