gpt-5.6-terra
—30-day success rate
Collecting data…
Average latency: —30-day calls: —
Official models · Stable and secure
Official GPT, Claude and Gemini models.
As low as 50% off.
flatkey.ai routes your traffic to the official GPT, Claude, and Gemini APIs through one key. Model prices run 60-90% of official list, and topping up $200 gets $100 free — stacked, as low as half the official price, stable and secure.
Two discounts, stacked
Input price per 1M tokens
- 1Model prices: 60-90% of official list
- 2Top up $200, get $100 free — you pay 2/3
OfficialAfter bonusgpt-5.6-terraOpenAI
claude-opus-4-8Anthropicclaude-sonnet-5Anthropic
gemini-3-pro-deepsearchGoogle
gpt-5.6-terraOpenAI$2.5
$1
claude-opus-4-8Anthropic$5
$2.6667
claude-sonnet-5Anthropic$2
$1.0667
gemini-3-pro-deepsearchGoogle$75
$40
Input / 1M tokens · Both discounts stack — as low as 50% of the official price
46Btokens served monthly
4K+paying users
45models behind one key
100+enterprises in production
Live model health
30-day health, measured on real traffic
Every number below comes from real production calls routed through flatkey.ai — success rate, latency, and volume over the last 30 days. No synthetic benchmarks.
claude-opus-4-8
—30-day success rate
Collecting data…
Average latency: —30-day calls: —
claude-sonnet-5
—30-day success rate
Collecting data…
Average latency: —30-day calls: —
Why teams pick flatkey.ai
Reliable, cheaper, and private by design
Stable and reliable
100+ enterprises and 4K+ paying users run on flatkey.ai every day.
- 99.9% average success rate over the last 30 days
- Automatic failover across multiple upstream providers
- Live health dashboards for every model
Cut your model spend
Model prices at 60-90% of official, and the top-up bonus takes another third off — as low as 50% of official. One key for 40+ models so every task runs on the right-cost model.
- 60-90% of official pricing, stacked with the top-up bonus — as low as 50% off
- 40+ models integrated behind one key
- Route cheap tasks to cheap models, hard tasks to frontier models
Privacy guaranteed
Our servers and storage follow GDPR, SOC 2, and ISO 27001 practices. We do not store your prompts or completions.
- GDPR compliant infrastructure
- SOC 2 and ISO 27001 aligned controls
- Zero retention of your request content
All models
45 models, one key — prices, latency, and health
The 10 busiest models: after-bonus price vs official, TTFT latency, and 30-day health. The full directory lives on the models page.
| Model | Official$ / 1M input tokens | After bonus$ / 1M input tokens | Latency | 30-day health |
|---|---|---|---|---|
| claude-haiku-4-5-20251001Anthropic | $1 | $0.533333 | — | — |
| claude-opus-4-5Anthropic | $5 | $3 | — | — |
| claude-opus-4-6Anthropic | $5 | $2.6667 | — | — |
| claude-opus-4-7Anthropic | $5 | $2.6667 | — | — |
| claude-opus-4-8Anthropic | $5 | $2.6667 | — | — |
| claude-sonnet-4-5Anthropic | $3 | $1.6 | — | — |
| claude-sonnet-4-6Anthropic | $3 | $1.6 | — | — |
| claude-sonnet-5Anthropic | $2 | $1.0667 | — | — |
 deepseek-ai/deepseek-r1DeepSeek | $75 | $40 | — | — |
| deepseek-ai/DeepSeek-V3DeepSeek | $0.25 | $0.133333 | — | — |
How it fits together
From homepage to production calls
1
Get one key
Create a flatkey account, open the dashboard, and generate an API key for your app.
2
Change the base URL
Point your OpenAI-compatible client to https://router.flatkey.ai/v1 and keep your existing SDK.
3
Monitor and optimize
Review usage, cost, routing, and errors from the same product dashboard.
Support
Questions? Talk to us.
We provide online support and usage consultation for the LLM services on flatkey.ai — integration, model choice, or billing. Reach us through any channel below.
Ready to replace
model chaos with one key?
Start from the flatkey homepage, manage your product dashboard, and keep router.flatkey.ai as the stable API endpoint.