fix: Add deepseek-v4.1-flash to the international Alibaba Token Plan provider

[github-actions]

Summary

Add missing deepseek-v4.1-flash on the international Alibaba Token Plan provider (alibaba-token-plan, Singapore). The model is already listed on Personal and Team Token Plan docs and exists on China Token Plan (alibaba-token-plan-cn via #7045); the lab metadata file is already present. This is an override-only provider row: base_model, reasoning controls, interleaved reasoning field, and zeroed subscription costs.

Changes

New file: providers/alibaba-token-plan/models/deepseek-v4.1-flash.toml

Field	Value
base_model	deepseek/deepseek-v4.1-flash
reasoning_options	toggle + effort low / high / max
interleaved.field	reasoning_content
cost	input/output/cache_read/cache_write = 0 (Credits subscription, same as sibling Token Plan rows)

Mirrors providers/alibaba-token-plan-cn/models/deepseek-v4.1-flash.toml with international source URLs.

Evidence

Claim	Source	What it establishes
Model on Token Plan Personal (Singapore)	https://www.alibabacloud.com/help/en/model-studio/token-plan-personal-overview	Supported models table lists deepseek-v4.1-flash (reasoning, visual understanding, text)
Model on Token Plan Team (Singapore)	https://www.alibabacloud.com/help/en/model-studio/token-plan-team-overview	Same model ID in Team supported models
Hybrid thinking + enable_thinking + reasoning_content	https://docs.qwencloud.com/api-reference/chat/openai-chat	DeepSeek-V4 series: enable_thinking; thinking in reasoning_content
DeepSeek on Alibaba / Singapore surface	https://www.alibabacloud.com/help/en/model-studio/deepseek-api	Documents deepseek-v4.1-flash, Singapore endpoints, thinking stream on reasoning_content
Lab metadata already exists	models/deepseek/deepseek-v4.1-flash.toml	Provider entry is override-only
China peer shape	providers/alibaba-token-plan-cn/models/deepseek-v4.1-flash.toml	Same base_model, toggle + low/high/max, interleaved, zero cost
Zero USD cost	Token Plan overview + existing alibaba-token-plan rows	Credits subscription; no public per-token USD rate on this endpoint

Validation

• Confirmed model ID on first-party Personal and Team Token Plan docs (accessed 2026-09-15).
• Confirmed lab file and China Token Plan peer exist; international provider lacked the row.
• Did not run bun validate or live API calls (no shell in this workflow).

Review notes

• QwenCloud OpenAI chat docs describe reasoning_effort for deepseek-v4.1-flash as integers 1–100, while the DeepSeek Model Studio page and the merged China entry use string efforts with effective ladder low/high/max. This PR keeps parity with China #7045 / Token Plan CN as requested; if Alibaba later documents a distinct control surface for international Token Plan only, update both providers together.
• Did not add pay-as-you-go alibaba international pricing for this model (out of issue scope).

Closes #7146

Automated by the issue fixer: https://github.com/anomalyco/models.dev/actions/runs/34942956189