Models
FreeGET /api/modelsThe /api/models endpoint returns the complete AI model catalog: per-model input price, output price, context window, capabilities, tier (flagship / mid / budget), and release date. The catalog is updated daily by merging community-maintained sources (LiteLLM) with our own curated baseline.
When to use this endpoint
When your agent needs the canonical pricing or specs for a model. For ranked recommendations based on a task and budget, use /api/premium/routing instead.
Example response
{
"ok": true,
"source": "tensorfeed.ai",
"lastUpdated": "2026-04-27",
"providers": [
{
"id": "anthropic",
"name": "Anthropic",
"models": [
{
"id": "claude-opus-4-7",
"name": "Claude Opus 4.7",
"inputPrice": 15,
"outputPrice": 75,
"contextWindow": 1000000,
"capabilities": ["text", "vision", "tool-use", "code"],
"tier": "flagship"
}
]
}
]
}Code samples
Python SDK
from tensorfeed import TensorFeed
tf = TensorFeed()
catalog = tf.models()
for provider in catalog["providers"]:
for m in provider["models"]:
print(f"{m['name']}: ${m['inputPrice']}/${m['outputPrice']} per 1M")TypeScript SDK
import { TensorFeed } from 'tensorfeed';
const tf = new TensorFeed();
const catalog = await tf.models();
const flagship = catalog.providers
.flatMap(p => p.models.filter(m => m.tier === 'flagship'));MCP tool
Available via the TensorFeed MCP server as get_model_pricing. Add npx -y @tensorfeed/mcp-server to your Claude Desktop or Claude Code MCP config.
FAQ
How often is the models catalog updated?
Daily, at 7am UTC. We merge LiteLLM's community-maintained pricing with our own curated baseline so new models from major providers land within a day.
Are the prices in /api/models always current?
Yes for the major providers we track. Edge cases (new providers, regional variants, volume discounts beyond the published rate cards) may lag. For production budgeting use /api/premium/cost/projection which projects against the live catalog.
What is the difference between /api/models and /api/pricing?
Same data. /api/pricing is a legacy alias kept for backwards compatibility. New integrations should use /api/models.