feat(ollama-cloud): add deepseek-v4-pro:0813 model

[YongYuH]

Context

Ollama Cloud has been serving deepseek-v4-pro:0813 (the 0813 snapshot of DeepSeek V4 Pro, listed by GET /v1/models since 2026-08-13). The lab metadata already exists on dev at deepseek-v4-pro-0813.toml; this PR adds the missing provider entry so the model surfaces under ollama-cloud in the generated catalog.

Confirmed live on the Ollama Cloud API:

$ curl -s https://ollama.com/v1/models | jq -r '.data[] | .id' | grep deepseek
deepseek-v4.1-flash
deepseek-v4-flash:0731
deepseek-v4-pro:0813

$ curl -s https://ollama.com/api/show -d '{"model":"deepseek-v4-pro:0813"}' | jq '.'
{
  "capabilities": ["completion", "tools", "thinking"],
  "details": {
    "parent_model": "deepseek-v4-pro:0813",
    "family": "deepseek4",
    "parameter_size": "1650497936906",
    "quantization_level": "FP8"
  },
  "model_info": {
    "deepseek4.context_length": 1048576,
    "general.architecture": "deepseek4"
  },
  "modified_at": "2026-08-13T08:00:00-07:00"
}

What changed

• Added providers/ollama-cloud/models/deepseek-v4-pro:0813.toml, a base_model-factored provider entry that inherits provider-agnostic facts from the existing lab metadata and only declares host-specific overrides.

Critical Changes

• New model on ollama-cloud: deepseek-v4-pro:0813 now resolves in the generated catalog with limit.context = 1_048_576 / limit.output = 1_048_576 and reasoning_options of toggle + effort: ["high", "max"].
• Override-only: the provider file uses base_model = "deepseek/deepseek-v4-pro-0813" and restates only reasoning_options and [limit] (real deltas vs. the lab's 1M/384K limits). name, description, family, dates, capabilities, modalities, open_weights, license, and weights are inherited from the lab entry.
• Reasoning controls: toggle + effort: ["high", "max"] matches the DeepSeek V4 family on this provider (see deepseek-v4-flash:0731.toml and deepseek-v4-pro.toml) and the AGENTS.md DeepSeek V4 guidance (toggle + high/max, not L/M/H). Wire path documented in the leading comment per the ollama-cloud convention (see feat(ollama-cloud): add deepseek-v4-flash:0731 model #3985).
• Context override rationale: the lab's 1_000_000 is a rounded figure; Ollama Cloud's /api/show exposes the real deepseek4.context_length = 1_048_576.
• limit.output = 1_048_576: Ollama publishes no separate output cap for this model. The same-host deepseek4 peers (deepseek-v4-flash:0731, deepseek-v4-pro) both serve output = 1048576, so this follows the established ollama-cloud convention (feat(ollama-cloud): add deepseek-v4-flash model #1586) rather than the lab's 384_000.
• No [interleaved]: Ollama's OpenAI-compatible endpoint returns the reasoning trace in message.reasoning (verified in ollama/ollama openai/openai.go), which is not a schema-supported interleaved field. The earlier deepseek-v4-pro entry on this provider had [interleaved] removed for the same reason (see 3a58678).

Expected behavior

• bun validate passes (exit 0).
• Resolved provider model:
  • id: deepseek-v4-pro:0813
  • name: DeepSeek V4 Pro 0813 (inherited from lab)
  • limit: { context: 1048576, output: 1048576 }
  • reasoning_options: [{ type: "toggle" }, { type: "effort", values: ["high", "max"] }]
  • All other fields inherited from models/deepseek/deepseek-v4-pro-0813.toml.

Relationship to #5109 and #6742

Both open PRs add the same file. This PR is a corrected alternative — happy to defer to the maintainers on which to land.

1. fix(ollama-cloud): correct 3 context_lengths, add 3 served-but-missing tags #5109 uses a full inline definition. Per AGENTS.md, non-lab hosts must use base_model when the lab model is nameable, and the lab metadata already exists. The automated review flagged this as a violation on that PR.
2. feat(ollama-cloud): add published token costs to all cloud models #6742 uses base_model correctly but bundles this file into a catalog-wide [cost] change. No other ollama-cloud model on dev currently has a [cost] block, so adding one only here would be inconsistent. This PR stays scoped to the missing model; if feat(ollama-cloud): add published token costs to all cloud models #6742 lands first, the cost block can be added there.

[github-actions]

No actionable findings.

[github-actions]

No actionable findings.