feat(cline-pass): add deepseek-v4.1-flash

[munim]

Closes #6861.

deepseek-v4.1-flash is available in ClinePass but missing from the provider data.

Adds providers/cline-pass/models/cline-pass/deepseek-v4.1-flash.toml as a full standalone entry, same shape as cline-pass/glm-5.3-flash.toml:

• model id cline-pass/deepseek-v4.1-flash (from the file path), name = "DeepSeek 4.1 Flash", description from the ClinePass catalog
• no base_model: base_model = "cline-pass/deepseek-v4.1-flash" does not resolve — base_model is looked up only in models/<lab>/<model>.toml (packages/core/src/generate.ts:161) and there is no models/cline-pass/ namespace, so the metadata is inlined instead
• metadata inlined from the DeepSeek V4.1 Flash lab entry: family deepseek-flash, 1M context / 384K output, text+image input, open weights, release 2026-09-10
• reasoning effort low|high|max + interleaved reasoning_content, matching the DeepSeek V4.1 Flash reference and the same-surface peer relay (opencode-go/deepseek-v4.1-flash)
• reference pricing = DeepSeek's off-peak V4.1 Flash rate, $0.15 / $0.60 / $0.003 per 1M tokens (input/output/cache read), the same rate models.dev records for the first-party DeepSeek entry; reasoning tokens bill at the output rate (reasoning = 0.6). Peak rates are $0.30 / $1.20 / $0.006.

Sources:

• ClinePass catalog: https://api.cline.bot/api/v1/ai/cline/recommended-models (accessed 2026-09-11)
• DeepSeek pricing: https://api-docs.deepseek.com/quick_start/pricing (accessed 2026-09-11)

bun validate passes.

[github-actions]

No actionable findings.

[github-actions]

No actionable findings.

[github-actions]

Action items

• [high] [violation] providers/cline-pass/models/cline-pass/deepseek-v4.1-flash.toml:11 - Check: Non-lab hosts must use base_model when the lab model is nameable; provider files stay override-only. Why: ClinePass did not create DeepSeek V4.1 Flash. models/deepseek/deepseek-v4.1-flash.toml already exists, and other hosts (including this PR’s first commit and peers like opencode-go) correctly use base_model = "deepseek/deepseek-v4.1-flash". The final full inline copy is not a unique-to-host exception; sibling ClinePass DeepSeek entries (deepseek-v4-flash, deepseek-v4-pro) already use base_model. Action: Restore base_model = "deepseek/deepseek-v4.1-flash" and keep only provider deltas (cost, reasoning_options, interleaved, and any real overrides such as ClinePass name/description). Remove restated identical lab fields (family, dates, booleans, knowledge, [limit], [modalities], etc.).