NewDeepSeek V4.1 Flash 1017, now livedeepseek-v4.1-flash-1017, now live

Every frontier model. One verifiable API.

Route Claude, GPT, Gemini, DeepSeek, Qwen and more through one OpenAI-compatible endpoint. Compare real latency, verify every call, and pay only for what you use.

Get API key

Your appclient.chat.completions.create( model="claude-opus-5-5", messages=[…])

api.quicksilverpro.io/v1model echoed · cost computed

Response · 200 OKsample call

usage.cost$0.0031401,240 → 380 tokens75% below list

✓ model echoed✓ matches /pricing.json

+ 86 more in the catalog

your appclaude-opus-5-5

Response · 200 OKsample call

$0.003140−75% vs list

✓ model echoed✓ matches /pricing.json

Start from the work, not the model list.

Pick a job and get the models built for it. Only the model string changes.

Curated from catalog data, not a benchmark.

Your codemodel="claude-opus-5-5"

$1.00 / $5.00Input / output per 1M

75% lowervs Anthropic list $4.00 / $20.00

Context: 1M
AA Index: 58
Status: Checking...

All coding models Full pricing

Keep your SDK. Change the base_url.

QuickSilver Pro speaks the OpenAI wire format, so existing clients, agent frameworks and the qsp CLI keep working.

Before · any OpenAI clientapp.py

client = OpenAI(
    base_url="https://api.openai.com/v1",
    api_key=os.environ["OPENAI_API_KEY"],
)
client.chat.completions.create(
    model="gpt-4o", …)

After · QuickSilver Proapp.py

client = OpenAI(
    base_url="https://api.quicksilverpro.io/v1",
    api_key=os.environ["QSP_KEY"],
)
client.chat.completions.create(
    model="claude-opus-5-5", …)

client = OpenAI(- base_url= - "https://api.openai.com/v1",+ base_url= + "https://api.quicksilverpro.io/v1",)model="claude-opus-5-5" # any of 91

base_url: - https://api.openai.com/v1; + https://api.quicksilverpro.io/v1
model: claude-opus-5-5 # any of 91

91 modelsCatalog v68
OpenAI-compatibleChat · Responses · Images · Speech
Live statusPer-model status
Published rates/pricing.json

Every call comes with proof.

Confirm the model returned, reconcile usage.cost against published rates, and check live status.

POST /v1/chat/completions · 200 OKsample call

{  "object": "chat.completion",  "model": "claude-opus-5-5",✓ echoed  "choices": [ … ],  "usage": {    "prompt_tokens": 1240,    "completion_tokens": 380,    "cost": 0.003140✓ matches /pricing.json  }}

GET /v1/status · claude-opus-5-5 ·checking

Trust & verification Live status

New models, live on day one.

A model that appeared, a price that moved. Every entry is dated and checkable.

Oct 27 · latestGPT-6.1 Terra and Terra ProOpenAI's GPT-6.1 Terra joins the catalog at OpenRouter list parity: $1 / $5 per million input / output tokens, cached input $0.10, cache write $1.25…
Oct 25 · launchBria FIBO 2.0Bria FIBO 2.0 is available through the OpenAI Images API at $0.055 per image, matching FIBO 1.5, which remains available.
Oct 23 · launchKimi K3.1 Pro and HighspeedTwo Moonshot routes at endpoint list parity: Kimi K3.1 Pro $2 / $10 and Highspeed $4 / $20 per million input / output tokens…

Full changelog

One API, from first call to private deployment.

Hosted API

QuickSilver Pro for teamsTeams

Production traffic on the same 91-model API, with invoice billing and net-30 terms, volume pricing past $5K/mo, a contracted uptime target, and dedicated per-key capacity.

Invoice billing, volume pricing past $5K/mo, a contracted uptime target.

For teams

Enterprise capability · separate deploymentEnterprise

Your own private gatewayPrivate gateway

One private endpoint for Claude Code, Codex and every agent tool, with central budgets and governance. It isn't part of the standard hosted API.

Self-hosted · air-gapped · private endpoint. A separate deployment, not part of the hosted API.

Self-hosted
Air-gapped
Private endpoint

Talk to Enterprise

Start with one call.

Start with $0.05 credit on starter models. No card.$0.05 starter credit · no card