01STANDARD
STARTER
10MTOKENS
For quick tests and setup
$5SELECT ↗
FABUS® / MODEL ROUTING INFRASTRUCTURE / 2026
ONEROUTINGFABRIC.
[ PRODUCT 001 ]
Buy model capacity once. Route every coding agent through a single OpenAI-compatible surface.
✦ CURSOR✦ OPENCODE✦ CODEX✦ CLINE✦ ZED✦ JETBRAINS✦ REST API
CAPACITY,
READY TO ROUTE.
Prepaid model tokens. No subscription. No provider juggling. Packs stay available until you use them.
02 / LIVE FABRICMANY MODELS.
MANY MODELS.
ONE ADDRESS.
Fabus resolves a public model name into the best available upstream route, checks capacity, and bills one normalized balance.
PUBLIC MODELRATESTATE
01fabus/gemini-3.1-pro6OUTAGE
02fabus/claude-fable-510OUTAGE
03fabus/gpt-5.51OUTAGE
04fabus/gemini-3.1-pro-preview6OUTAGE
05fabus/claude-opus-512OUTAGE
06fabus/gpt-5.6-luna2OUTAGE
07fabus/gpt-5.6-sol2OUTAGE
08fabus/claude-sonnet-58OUTAGE
ONE KEY.
HARD LIMITS.
Issue isolated API keys, set daily and monthly caps, revoke access instantly.
EVERY TOKEN.
VISIBLE.
Audit input, output and cached tokens by key, model, hour and request.
DROP-IN
OPENAI API.
Move existing clients by changing one base URL and one model name.
bash
curl https://api.fabus.fun/v1/chat/completions \
-H "Authorization: Bearer fbs_live_..." \
-H "Content-Type: application/json" \
-d '{
"model": "fabus/claude-sonnet-4-6",
"messages": [{"role":"user","content":"Review this diff"}],
"stream": true
}'