Agent orchestration · hundreds of AI models
Harness orchestration · Nexus routing · session budgets & compaction · transparent channel pricing
ModelAPI is an agent & multi-model orchestration cloud — routing, budgets, compaction, and cache. You pay for orchestration software with tiered AKL / Rsc / Bedrock / direct channel pricing.
Connecting you to the world's leading AI providers
Trending This Week
Models ranked by usage
Based on real request volume over the past 7 days · updated hourly
Claude Code + server-side hybrid routing~$2.20/M output in typical agents
Harness sends tests, lint, and file scans to nexus/cheapest-west while plan/reasoning stays on frontier tiers. Verify via X-Harness-Task — see /harness.
Sound familiar?
3 blockers that stop most developers
Account friction, cross-border routing, and billing — we offer a compliant SaaS path (see /compliance).
Account rejected or banned
Overseas phone + credit card required. Accounts banned without warning; balance lost.
✓ Our solution
Sign up with email. Eligible users per /compliance; credits never expire.
Unstable network & timeouts
Cross-border API calls over long routes suffer latency spikes and timeouts — hard to meet production SLAs.
✓ Our solution
Multi-region PoPs and smart routing. Primary endpoint in US East; regional BASE_URLs as PoPs ship.
Payment walls & no invoices
Foreign card only. No RMB payment, no VAT invoice, no corporate billing.
✓ Our solution
WeChat Pay, Alipay, bank transfer in RMB. VAT invoices available. Corporate billing supported.
Problems we solve
Ten keys, inflated bills, flaky routing — condensed into one control plane.
10 provider accounts · 10 API keys · 10 dashboards
OpenAI, Anthropic, Google, DeepSeek — each with its own account, billing, key rotation, and quota tracking.
API feels 10× more expensive · hidden token inflation
Per-token API billing adds up. Some gateways add markup, inflate counts, or swap models.
Cheap models underperform · flagship costs · manual switching hurts flow
Trading cost vs quality manually breaks unattended workflows.
Complex docs · hard to debug · slow support
Non-developers get stuck; ticket cycles are long.
Frontier paths and savings in one envelope
Same credential surface for experimentation and scaled traffic.
Frontier models
Claude Fable 5 / Sonnet 5, GPT-5.6 Sol/Terra/Luna, Opus 4.8, Gemini 3.1 Pro — same integration surface.
High-value models
DeepSeek V4 Flash, Qwen, Kimi, GLM — built for volume and everyday workloads.
Watch a request end-to-end
Representative routing: code pane, streamed body, latency and credits.
Loading...Ecosystem
Works where you already build
Point Cursor, Windsurf, VS Code, Claude Code, or Codex CLI at ModelAPI.
Universal config — works in every tool
Resilience
Backup hosts when routes change
Reuse the exact same key — swap hostname only.
Catalog
Featured models
50+ frontier and value tiers · one unified API envelope
Claude Sonnet 5
GPT-5.6 Sol
Claude Opus 4.8
GPT-5.6 Terra
GPT-5.6 Luna
Claude Fable 5
Gemini 3.1 Pro
GPT-5.5
Qwen3.7 Max
GLM-5.1
Kimi K2.7 Code
DeepSeek V4 Flash
Capabilities
What sits behind one endpoint
Harness, Nexus, and failover algorithms — one endpoint, not another passthrough shop.
Harness Control
Server-side agent orchestration: tiered routing, session budgets, auto-compaction, and cache tuning via HTTP headers — proprietary algorithm layer.
Nexus Smart Routing
nexus/* aliases resolve to the best model by cost, latency, or capability. Software routing with fallbacks — not blind passthrough.
Universal Model Access
GPT-5.5, Claude Opus 4.8 / 4.7, Gemini 3.1, DeepSeek V4, Grok 4.3 and 50+ more — all through one API key, no restrictions.
Drop-in OpenAI SDK
Change 2 lines of code. Keep your existing SDK, prompts, and tools — zero refactoring.
Global Edge Network
US · Singapore · Tokyo · Frankfurt. Regional PoPs roll out after US primary launch.
Lowest Total Cost
Official API prices, zero markup. 5% infrastructure fee at top-up only — below industry average. Bring Your Own Key (BYOK) at $0.50/M for expensive models.
Enterprise Security
Key isolation · RBAC · Audit logs · Spend caps · SOC 2 compliant infrastructure.
Agent cloud · not a token relay
Harness Control: orchestration for agents
ModelAPI sells an agent service: tiered task routing, session budgets, compaction, and cache optimization. Model calls are billed at upstream channel rates — you buy orchestration, not a simple token hop.
Harness Control
Server-side agent orchestration: tiered task routing (plan / bulk / compress), session USD budgets, auto context compaction, and prompt-cache optimization.
Nexus Routing Engine
Cost-, latency-, and capability-aware model selection via nexus/* aliases. Live catalog resolution — your code stays stable as frontier SKUs update.
Multi-tier Fallback
Profile-driven fallbacks (e.g. Opus → Sonnet → nexus/cheapest-west) with provider health checks — intelligent agent routing.
Response cache & Memory API
Exact-match response cache for repeated non-stream calls; transient memory consolidate API for Hermes/OpenClaw agent workflows.
curl https://api.aimodelapi.ai/v1/chat/completions \
-H "Authorization: Bearer sk-ama-..." \
-H "x-harness-profile: claude-code-optimal" \
-H "x-harness-session-id: my-project-1" \
-H "x-harness-auto-compact: true" \
-d '{"model":"nexus/best-coding","messages":[...]}'Response headers X-Harness-Profile / X-Harness-Model / X-Harness-Task expose routing decisions. See /docs#harness and integrations.
For Claude Code power users
What Anthropic direct cannot do — Harness does server-side
Claude Code is a great terminal agent. ModelAPI Harness is orchestration software on our gateway — it does not replace Claude Code, but auto-picks models, enforces budgets, and compacts long context. All features below are live and verifiable via response headers.
You keep Claude Code
Set ANTHROPIC_BASE_URL to api.aimodelapi.ai and add x-harness-profile: claude-code-optimal.
Harness reads each request
"Run all tests" → bulk tier → nexus/cheapest-west. Architecture question → plan tier → nexus/best-reasoning.
You see the decision
X-Harness-Task and X-Harness-Model tell you which model ran — auditable, billable at official rates.
| Capability | Anthropic direct | Claude Code + Harness |
|---|---|---|
Multi-vendor models in one agent session Opus for plan, DeepSeek Flash for bulk — automatic via tiers | No | Yes |
Server-side task tier routing (plan / bulk / compress) Rules + optional hybrid LLM classifier on every request | No | Yes |
Session USD spend cap with HTTP 429 Redis-backed; claude-code-optimal default $50/session | No | Yes |
Auto context compaction before expensive calls Server summarizes older turns; non-streaming requests today | Partial | Yes |
Live model aliases (nexus/*, claude-sonnet-latest) | No | Yes |
Provider prompt-cache optimization (Bedrock) Dashboard shows cacheRead / cacheSavingsUsd | Partial | Yes |
Exact-match response cache Non-stream, temp ≤ 0.1, no tools; X-ModelAPI-Cache header | No | Yes |
Agent memory consolidate API POST /v1/harness/memory/consolidate | No | Yes |
Transparent routing in response headers X-Harness-Model, X-Harness-Task, X-Harness-Classifier | No | Yes |
Native Claude Code UX (queue, steer, IDE) Keep Claude Code — point ANTHROPIC_BASE_URL at ModelAPI + Harness headers | Yes | Partial |
Honest notes: Harness does not change Claude Code's client UI. Auto-compact and response cache apply mainly to non-streaming calls today. "Response cache" is exact payload match, not embedding similarity. Hybrid cost ($2.20/M output at 5% Opus / 95% cheap) is a typical agent workload estimate — your mix may differ.
Start in under 60 seconds
Three steps to your first call
Add credits
Top up from $5. Supports WeChat, Alipay, Stripe, Apple Pay, crypto.
View payment optionsGet your API key
Create a key, swap the base URL. Your existing OpenAI code works as-is.
Read quickstartPricing
Usage-based tiers without guesswork
Upstream model passes through with a spelled-out infra surcharge — zero surprise fees.
Free
For exploration and testing
- 20 req/min · 50K req/day
- 10+ free models
- Community support
- Basic analytics
Pay-as-you-go
Official API prices · zero markup · min $5 top-up
- 500 req/min · 50K req/day
- All 50+ models
- 0% per-token markup
- 5% infra fee (below industry avg)
- Bring Your Own Key · $0.50/M
- Navigator 1 AI support · 7×24
- WeChat · Alipay · Stripe · Apple Pay
Enterprise
For production workloads
- Unlimited rate limits
- Reduced service fee (negotiated)
- Dedicated infrastructure
- SLA guarantee
- 24/7 dedicated support
- VAT invoice available
Accepted payment methods worldwide
Loved by developers
Real users, real results
"Used to juggle 7–8 API accounts. Now one ModelAPI key covers everything, billing is crystal clear, and switching between Claude Opus and DeepSeek takes zero effort."
"The Singapore PoP is legitimately fast. I was skeptical about a gateway adding latency, but median response time is actually lower than hitting the OpenAI API directly from SEA. Solid routing."
"We needed one gateway for multiple model providers. ModelAPI unified billing and routing; WeChat Pay works for our eligible accounts. Deployed to production in one afternoon."
"The PPP pricing for India is real — I'm paying ~40% less than teammates in the US for the same models. Switching my existing codebase took literally 2 lines of code."
"Looked at a lot of gateways — ModelAPI is one of the few with truly 0% per-token markup and a transparent fee structure. The nexus/cheapest route cut our batch inference cost substantially."
"Building multi-tenant where users have different model preferences. The Compare page saved hours of manual benchmarking. BYOK at $0.50/M flat is the most cost-effective routing I've found."
Recent Updates
What we've shipped lately
Claude Opus 4.8 — live via ArkLin (OEM + AKL)
Use claude-opus-latest or AKL-claude-opus-4-8. Save cost with AKL-*-rsc. Hermes dropdown "Opus 4 7" may fail — see docs.
DeepSeek V4 Flash added — 344% weekly surge
Now routing via nexus/cheapest. $0.14/M input, 1M context.
Human-in-the-loop tool calls now supported
Pause agentic workflows mid-execution and inject human decisions via the API.
nexus/cheapest-west routing now 40% faster with edge caching
Prompt-level cache hits on repeated context now reduce latency by ~40% at no extra cost.
Access every top AI. Right now.
Sign up in 30 seconds. No credit card required. 50+ models, one key.