One key.
Every model.
A gateway that resells one quota across eight model families. Point any OpenAI or Anthropic client at /v1 and stop juggling accounts.
| t+ | model | p50 |
|---|---|---|
| 00:00:04 | claude-sonnet | 412 |
| 00:00:12 | deepseek-v4 | 288 |
| 00:00:19 | gemini-3-pro | 531 |
| 00:00:26 | gpt-5.2 | 604 |
| 00:00:33 | kimi-k2 | 352 |
Model catalog
Pricing you can read before you run.
Multiplier, context window, throughput and p50 latency for every model on the gateway. Sort it, filter it, then point your client at the one you picked.
| Model | Pricing | Modality | Context | Throughput | Latency | 24h status | ID |
|---|---|---|---|---|---|---|---|
| Abliteration FP81 model | |||||||
ablit-fp8-70b | 0.25× | text | 64K | 220 t/s | 240ms | live | ablit-fp8-70b |
| Claude2 models | |||||||
claude-haiku-4.5 | 1× | text · vision | 200K | 132 t/s | 410ms | live | claude-haiku-4.5 |
claude-sonnet-4.5 | 3× | text · vision | 200K | 78 t/s | 640ms | live | claude-sonnet-4.5 |
| DeepSeek3 models | |||||||
deepseek-coder-v3 | 0.6× | text | 128K | 70 t/s | 810ms | live | deepseek-coder-v3 |
deepseek-v3.2 | 0.5× | text | 128K | 64 t/s | 880ms | live | deepseek-v3.2 |
| Gemini2 models | |||||||
gemini-2.5-flash | 0.5× | text · vision | 1M | 210 t/s | 260ms | live | gemini-2.5-flash |
gemini-2.5-pro | 2× | text · vision | 1M | 88 t/s | 720ms | live | gemini-2.5-pro |
| GLM (Zhipu)2 models | |||||||
glm-4.5-air | 0.4× | text | 128K | 110 t/s | 480ms | live | glm-4.5-air |
glm-4.6 | 1× | text · vision | 200K | 58 t/s | 900ms | live | glm-4.6 |
| Kimi (Moonshot)2 models | |||||||
kimi-k2-turbo | 1.8× | text | 256K | 148 t/s | 520ms | live | kimi-k2-turbo |
kimi-k2 | 1.2× | text | 256K | 52 t/s | 1010ms | live | kimi-k2 |
| MiMo (Xiaomi)1 model | |||||||
mimo-7b-rl | 0.3× | text | 128K | 190 t/s | 300ms | live | mimo-7b-rl |
| OpenAI1 model | |||||||
gpt-5.1 | 2.5× | text · vision | 400K | 96 t/s | 590ms | live | gpt-5.1 |
| Qwen1 model | |||||||
qwen3-max | 1.4× | text · vision | 256K | 74 t/s | 690ms | live | qwen3-max |
multiplier = tokens of quota spent per token used · latency measured p50 over 24h
How it works
Three steps, then it is just an API call.
Create an account
Email or a provider — the free tier lands with 10M tokens and one live key, no card.
Point your client
Swap the base URL to /v1 and drop the key in. OpenAI and Anthropic SDKs both work untouched.
Watch one quota
Every model draws from the same balance. Usage, latency and per-key spend show up live.
One key covers every model — no per-provider billing, no seat pricing.
curl /v1/chat/completionsWhat a key gets you
Built like infrastructure, priced like a utility.
Two protocols, one key.
OpenAI clients call /v1/chat/completions. Anthropic clients call /v1/messages. The same key and the same quota cover both, so switching tools means switching a base URL.
base_url = "https://xrouters.xyz/v1"Multipliers in the open.
Every model declares the weight it charges against your quota, and the number sits in the public model list. You can price a workload before you run it, not after.
deepseek-v3.2 · 0.5× · 128K ctxA dedicated endpoint for heavy use.
Enterprise keys get their own subdomain with raw provider multipliers, separated from shared traffic so a busy hour elsewhere does not become your latency.
https://<you>.vip.xrouters.xyz/v1Integration
Three lines to change.
There is no SDK to learn and no request shape to relearn. Keep the client you already use, swap the base URL, and the gateway handles protocol translation, quota accounting and retries upstream.
- Streaming passthrough — no buffering on the gateway
- Tool calls and multimodal parts survive the hop intact
- Quota is charged from the usage block the provider returns
- 402 with a readable message instead of a silent downgrade
curl https://xrouters.xyz/v1/chat/completions \
-H "Authorization: Bearer $XROUTERS_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v3.2",
"messages": [{"role":"user","content":"hello"}],
"stream": true
}'Pricing
Cheaper per token as you scale.
Pay for tokens, not seats. Top up from $1.50 in USDT, activate quota from your dashboard wallet instantly, and keep your tokens with no expiry.
Starter
$1.50 one-time · 100M tokens
For students, side projects, and initial prompt testing.
- OpenAI and Anthropic compatible
- All 28 models, public multipliers
- Usage dashboard with 24h history
Pro
$7.00 one-time · 500M tokens
For researchers running multi-model evaluation sweeps.
- Everything in Starter
- Token saver optimizer
- Per-key spend caps and alerts
Ultimate
popular$13.00 one-time · 1B tokens
For startups shipping production features without seat fees.
- Everything in Pro
- Priority routing queue
- Streaming + tool-call passthrough
Enterprise
$110.00 one-time · 10B tokens
For heavy traffic that needs an isolated VIP subdomain.
- Dedicated VIP subdomain
- Raw provider multipliers
- Invoice billing and SLA
one-time top-up · quota never expires · enterprise adds raw provider multipliers and a dedicated subdomain
Dashboard
The numbers, without the sales call.
Every key you hold, what it spent today, which models it touched, and the requests behind each charge. Admin gets the same view one level up — traffic by IP, error rate, host CPU, RAM and disk.
Questions
Answered plainly.
Still stuck after reading these? Write to support@xrouters.xyz with the key prefix and the request id — both are on every log row.
Create an account, open the dashboard, and press Create key. The key is shown once — copy it into your client's api_key field. You can hold several keys at once and revoke any of them without touching the others.
Anything that speaks the OpenAI chat-completions protocol or the Anthropic messages protocol: the official SDKs, LangChain, LlamaIndex, Cursor, Cline, Continue, LibreChat, Open WebUI, SillyTavern. You change one base URL — the rest of your code stays exactly as it is.
Every response reports its prompt and completion tokens. That count is multiplied by the model's public multiplier — a 2× model spends two tokens of quota per token used. Multipliers sit in the model list, so you can price a workload before running it.
Requests stop with a 402 and a clear message rather than silently degrading to a cheaper model. Top up from the dashboard and the same key resumes immediately — no reissue, no config change.
USDT on TRC-20 or ERC-20. The dashboard generates a reference code and an address; paste the transaction hash and the balance lands after confirmation. Minimum top-up is $1.50 and tokens never expire.
Yes. The model list carries live throughput and p50 latency sampled over 24 hours, plus a per-model status column. The admin console adds error rate, per-IP traffic and the raw request log.
Ready when you are
Ten million tokens to find out if this fits.
No card, no seat pricing, no expiring quota. Register, take the key, ship one request — then decide.
Voucher codes and referral credit both land in the same balance.
Running your own router? Peer it in with an ingress key and pass traffic through.