Models

CheapestInference serves six open-source models across three pools. A subscription is per-pool: during your reserved time blocks you get unlimited usage of every model in that pool — there is no separate full-catalog tier. Rate limits are set at the key level, not per model.

Flagship Pool — Kimi K3 (Moonshot’s flagship, from $126.65/mo with annual billing — very limited seats)
Frontier Pool — Kimi K2.7, GLM 5.2, MiniMax M3 (frontier coding & agentic models, from $48.45/mo with annual billing)
Core Pool — DeepSeek V4 Flash, MiMo v2.5 (fast 1M-context models, from $12.74/mo with annual billing)

List models

Query the live, authoritative model list:

curl https://api.cheapestinference.com/v1/models \
  -H "Authorization: Bearer YOUR_API_KEY"

Each model object includes an id, owned_by, and a type field so you can filter programmatically:

{
  "id": "kimi-k2.7",
  "object": "model",
  "created": 1677610602,
  "owned_by": "cheapestinference",
  "type": "chat"
}

Flagship Pool

Model	Model ID	Context	Cost basis (in / out per 1M)
Kimi K3	`kimi-k3`	256K	$3.00 / $15.00

From $126.65/mo with annual billing ($149/mo monthly). Moonshot’s current flagship, served solo in its own pool — very limited seats.

Frontier Pool

Model	Model ID	Context	Cost basis (in / out per 1M)
Kimi K2.7	`kimi-k2.7`	256K	$0.45 / $2.25
GLM 5.2	`glm-5.2`	198K	$1.40 / $4.40
MiniMax M3	`minimax-m3`	1M	$0.60 / $2.40

From $48.45/mo with annual billing ($57/mo monthly). Kimi K2.7 is Moonshot’s flagship; its predecessor Kimi K2.6 was retired from the pool in July 2026.

Core Pool

Model	Model ID	Context	Cost basis (in / out per 1M)
DeepSeek V4 Flash	`deepseek-v4-flash`	1M	$0.14 / $0.28
MiMo v2.5	`mimo-v2.5`	~1M	$0.14 / $0.28

From $12.74/mo with annual billing ($14.99–19.99/mo monthly depending on the block).

The “cost basis” is our underlying per-token cost, useful for comparison. On a time-block subscription you pay a flat monthly fee, not per-token charges. See Plans & Limits.

The set of models in a pool can change over time, and more pools may open. GET /v1/models is always the authoritative live list.

Per-model details:

Kimi K3 API — Moonshot’s flagship, served in the Flagship Pool
Kimi K2.7 API — Moonshot’s frontier agentic/coding model
Kimi K2.6 — Moonshot’s previous flagship, retired July 2026
GLM 5.2 API — Zhipu’s coding & reasoning model
MiniMax M3 API — frontier multimodal coding model with 1M context
DeepSeek V4 Flash API — DeepSeek’s fast 1M-context model
MiMo v2.5 API — Xiaomi’s fast, efficient ~1M-context model

Using models

Specify the model ID in your request:

# OpenAI SDK
response = client.chat.completions.create(
    model="kimi-k2.7",  # or "glm-5.2", "minimax-m3", "deepseek-v4-flash", ...
    messages=[{"role": "user", "content": "Hello"}]
)

All models work through the OpenAI endpoint (/v1/chat/completions) and the Anthropic-compatible endpoint (/anthropic/v1/messages). The API handles format translation automatically. Your key serves the models of the pool you subscribed to.

Context windows

Each model has a maximum context window (shown in the pool tables above). Requests that exceed it are rejected with HTTP 400:

Anthropic-dialect endpoints return {"type":"error","error":{"type":"invalid_request_error","message":"prompt is too long: <tokens> tokens > <limit> maximum"}}.
OpenAI-dialect endpoints return error.code = "context_length_exceeded".

When the model’s context window is known, responses include the x-model-context-tokens header with its value in tokens, so clients can read the exact limit for the model they are calling.

Claude Code tip: Claude Code can compact a session automatically before it fills the window. For a smooth fit, set CLAUDE_CODE_AUTO_COMPACT_WINDOW to the model’s context window — e.g. CLAUDE_CODE_AUTO_COMPACT_WINDOW=256000 for Kimi K3’s 256K window — so long sessions stay comfortably within the model’s context and keep flowing.