[nicho810]

All Submissions:

• Notify authors: * @agentforce314, @ericleepi314

New Feature Submissions:

• Have you added an explanation of what your changes do and why you'd like us to include them?

Description

Adds cheapaimodel (https://cheapaimodel.ai) as a row in the OpenAI-compatible ProviderSpec registry. It is an OpenAI-compatible model API gateway serving /v1/chat/completions (streaming supported) and /v1/models with Bearer auth, so it needs no bespoke class. Prepaid USD balance, billed per token. Docs: https://cheapaimodel.ai/docs. Model catalog: https://cheapaimodel.ai/models. API keys: CHEAPAIMODEL_API_KEY, created in the console after signing in at https://cheapaimodel.ai. Operated by LINKOVA PTE. LTD. (Singapore).

Disclosure: I work on cheapaimodel. This follows the same pattern as #972 (CloudGPU) and #971 (Runcrate).

Changes proposed in this pull request:

• src/providers/openai_compatible_specs.py: cheapaimodel spec after cloudgpu, base URL https://api.cheapaimodel.ai/v1, default model deepseek-v4-1-flash, five suggested models (deepseek-v4-1-flash, glm-5.3, glm-5.3-flash, kimi-k3, grok-4.7), CHEAPAIMODEL_API_KEY, aliases cheap-ai-model / cheap_ai_model. The comment notes that kimi-k3 only accepts temperature 0.6 or 1 and top_p 0.95.
• tests/test_provider_registry.py: cheapaimodel added to EXPECTED_NEW_PROVIDERS and VENDOR_DEFAULTS
• README provider list (and the translated READMEs): add cheapaimodel after cloudgpu. English count 32 to 33, translated counts 27 to 28.

Testing

• python -m pytest -q tests/test_provider_registry.py: 27 passed
• python -m pytest -q tests -k 'provider or registry or login or config': 1324 passed, 5 skipped (same as on main)
• get_provider_class('cheapaimodel')(api_key=...) resolves to CheapaimodelProvider with base URL https://api.cheapaimodel.ai/v1 and model deepseek-v4-1-flash; aliases cheap-ai-model and cheap_ai_model resolve to the same class.
• Unauthenticated GET https://api.cheapaimodel.ai/v1/models returns 401, so the host and path are live. I did not make authenticated model calls for this PR.

[agentforce314]

LGTM