Add MC-1 by ColomboAI

[wilkont]

Add MC-1 by ColomboAI

Summary

Adds the native mc1 provider and its logical colomboai/mc-1 model so
models.dev consumers such as OpenCode can discover MC-1 without a custom
provider definition.

Contract

• OpenAI-compatible Chat Completions base: https://api.colomboai.com/v1
• API-key environment variable: MC1_API_KEY
• text input and output
• tool calling and structured output
• 128,000-token context window
• 128,000-token maximum output with a 4,096-token default

The first public MC-1 repository commit containing the logical model is dated
2026-08-15: https://github.com/ColomboAI-com/MC-1-platform/commit/7d93ac4bdc854858b1e2d9fd2bb835bf3b536216

Pricing is omitted because MC-1 dynamically selects an eligible executable
route and does not publish one fixed per-token price for the logical model.

Validation

• bun validate passes against the current models.dev dev branch.
• MC-1's repository suite passes with runtime rejection, model-catalog,
  OpenAPI, documentation, and integration-example coverage for the fixed output
  limit.

[github-actions]

No actionable findings.

[wilkont]

Correction: the earlier 4,096-token maximum was superseded. MC-1 now has an enforced 128,000-token logical-model maximum and retains 4,096 only as the default when callers omit an output budget. Runtime route eligibility now excludes model profiles that cannot execute the requested budget. The authoritative product change and 563/563 validation are in https://github.com/ColomboAI-com/MC-1-platform/pull/40. The contribution metadata and PR body are aligned to 128,000.

[github-actions]

Action items

• [medium] [possible mistake] providers/mc1/models/colomboai/mc-1.toml:16 - Check: Provider limit.output must match the host’s documented maximum output. Why: The PR body and earlier commit set a fixed 4,096-token max output and claim MC-1 suite coverage for that fixed limit, but the final TOML sets output = 128000 (equal to full context) with no updated citation for the raise. Action: Confirm the real MC-1 max output against first-party docs/OpenAPI; set limit.output to that value and align the PR body (and any cited validation) with it.
• [low] [violation] providers/mc1/models/colomboai/mc-1.toml:14 - Check: Source/rationale comments belong in a leading header block above the first key. Why: The dynamic-pricing note sits mid-file under open_weights, so it is not in the required leading comment position. Action: Move the cost-omission rationale (and any docs/API citations) to a top-of-file comment block above name.

[github-actions]

No actionable findings.