OpenRouter alternatives, 2026.Q3: every real option, ranked
The short answer
Edition v2026.Q3 · pricing and status verified 2026-08-06.
LiteLLM is the strongest OpenRouter alternative for most — teams that need the router inside their own VPC — keys, logs, and spend controls that never leave your infrastructure. Then: Vercel AI Gateway · Cloudflare AI Gateway · Portkey. Below, all 19 real options in the model router category, with pricing and honest watch-outs — plus the 5 "alternatives" other lists still recommend that are dead, renamed, or sunsetting.
Why people look past OpenRouter at all: That 5.5% compounds painfully at volume — heavy spenders graduate to direct provider contracts or BYOK (which itself costs 5% past 1M req/mo). Fully managed only: your entire model traffic transits a single startup, and critics argue frontier-lab consolidation erodes the aggregator's reason to exist.
The top OpenRouter alternatives, ranked
2LiteLLMBerriAI (YC W23; MIT OSS)
Open source free forever (self-host) · Enterprise custom annual, sized to request capacity — never per-token · 30-day trialBest for Teams that need the router inside their own VPC — keys, logs, and spend controls that never leave your infrastructure.
Watch You run it — upgrades, scaling, and a famously sprawling codebase (2.5k open PRs) are your problem. Performance numbers are vendor-published; rivals (Bifrost) built entire marketing campaigns on LiteLLM's Python-era overhead. Enterprise pricing is quote-only and opaque.
3Vercel AI GatewayVercel
0% markup, 0% platform fee — provider list price via credits · BYOK free · free tier (subset of models) · metered add-ons (team-wide ZDR/allowlist $0.10/1k req)Best for TypeScript/AI SDK teams and anyone who refuses to pay a take rate on tokens at all.
Watch Youngest of the leaders; observability and governance arrive as metered add-ons (trace drains, tag writes, query fees) that quietly rebuild a bill. Loss-leader economics could change when strategy does. Deepest value only lands if you're inside the Vercel/AI SDK ecosystem.
4Cloudflare AI GatewayCloudflare
Core features free (analytics, caching, rate limits, dynamic routing) · unified billing +5% on credit purchases · Workers Paid ($5/mo) lifts log caps to 10M/gatewayBest for Teams already on Cloudflare who want a free, edge-speed control plane — caching, rate limiting, DLP — in front of any model API.
Watch Routing intelligence is thinner than dedicated routers — no learned model selection, and provider translation coverage trails OpenRouter's catalog. Observability depth is Cloudflare-grade metrics, not LLM-native evals. Least compelling if you're not otherwise a Cloudflare shop.
5PortkeyPortkey → Palo Alto Networks (closed May 29, 2026)
Gateway fully open source (Mar 2026) · Developer free (10k logs/mo) · Production $49/mo (100k logs, +$9/100k) · Enterprise custom (VPC, SOC2/HIPAA)Best for Enterprises that treat the gateway as a governance layer — guardrails, RBAC, budgets, prompt management, audit — not just a switchboard.
Watch The acquisition cuts both ways: the roadmap now serves Prisma AIRS, and the standalone product's independence is over. Buying it increasingly means buying Palo Alto. Pre-acquisition Portkey was also sub-scale commercially ($15M Series A) relative to its traffic.
Every other live option in model router
| Tool | Maker | What it is | Entry |
|---|---|---|---|
| Requesty | Requesty (Amsterdam) | EU-flavored OpenRouter alternative: 600+ models, 5% all-in markup, <14ms failover, 225B+ tokens/day claimed; $3M seed led by 20VC (Sep 2025) | 5% markup |
| Helicone AI Gateway | Helicone (YC W23) | Open-source Rust gateway from the observability company; cloud version with 0%-markup passthrough billing launched to beta Sep 2025 | free + tokens |
| Bifrost | Maxim AI | Go gateway marketing itself on speed — '50x faster than LiteLLM', <15µs overhead at 5k RPS (vendor benchmark), 1,000+ models, MCP built in | free + tokens |
| Kong AI Gateway | Kong Inc. | The API-gateway incumbent's AI layer: semantic routing/caching, prompt firewalling, multi-LLM plugins on Kong Gateway/Konnect | OSS free · Konnect tiers |
| Microsoft Foundry Model Router | Microsoft | Hyperscaler-native learned router, GA Nov 2025 (version 2025-11-18): picks among 28 models incl. GPT-5.x, Claude, DeepSeek, Grok inside Azure | Azure usage |
| Bedrock Intelligent Prompt Routing | AWS | Per-prompt routing between models of a family inside Bedrock; AWS-only answer to the router question | AWS usage |
| LLM Gateway | LLMGateway (OSS) | Open-source OpenRouter alternative; 0% fee with own keys, 5% on credits — publishes the category's best fee-comparison research | free + tokens |
| TrueFoundry AI Gateway | TrueFoundry | Enterprise self-hosted gateway (K8s-native); prolific comparison-content marketer aiming at LiteLLM/Portkey buyers | custom |
| Eden AI | Eden AI | Unified API beyond LLMs (vision, OCR, speech); 5.5% fee on credits — broader but shallower than LLM-first routers | credits + 5.5% |
| Not Diamond | Not Diamond | Trained per-prompt quality router (pick the best model, not just the cheapest); shipped a coding-agent router in 2026 — niche but alive | free tier + usage |
| Arch | Katanemo | Open-source agent-native proxy with its own Arch-Router model for preference-based routing | free + tokens |
| Apache APISIX AI Gateway | Apache Software Foundation | AI plugins (ai-proxy, rate-limiting, prompt guard) on the APISIX API gateway; solid if APISIX is already your edge | free |
| Higress | Alibaba (OSS) | Envoy-based cloud-native AI gateway, big in China deployments; token-aware routing and model failover | free |
| new-api | QuantumNous (OSS) | Popular Chinese-community multi-provider aggregator/reseller panel in the One API lineage | free + tokens |
| Glama Gateway | Glama | Low-fee gateway attached to an MCP-centric workspace; small but liked by indie agent builders | usage-based |
The "OpenRouter alternatives" to avoid — no longer what they were
Listicles still recommend these. As of 2026-08-06, they are not what the listicles think.
| Tool | Status | What happened |
|---|---|---|
| RouteLLM | fading | The academic cost-quality router that popularized learned routing (2024); repo largely quiet since — ideas absorbed by commercial routers |
| Martian | fading | Invented the 'model router' pitch ($9M, NEA/Prosus 2023; Accenture 2024) — has pivoted to AI interpretability research; router no longer the product |
| Unify | dead | Benchmark-driven LLM router (a16z-backed, 2024) — router retired; company pivoted to 'AI teammates' workflow agents by 2026 |
| TensorZero | dead | Rust gateway + LLMOps stack; archived its repo and shut down Jun 12, 2026, returning remaining capital — the category's starkest OSS casualty |
| MLflow AI Gateway | fading | Deployments/gateway module inside MLflow; maintained but overshadowed — Databricks invests via Mosaic AI Gateway instead |
How to choose
- If you just hit two models in production and have no infra team
- OpenRouter. One key, one invoice, ~25ms overhead; treat the 5.5% credit fee as rent until your spend justifies direct contracts.
- If keys, prompts, or logs cannot leave your VPC — or procurement says self-host
- LiteLLM open source. It's the de facto standard (Netflix, NVIDIA testimonials), and the Rust core removed the old performance excuse.
- If gateway fees offend you and you're in the TypeScript/AI SDK world
- Vercel AI Gateway — 0% markup and free BYOK is the best sticker price in the category; just watch the metered observability add-ons.
- If you're already behind Cloudflare and mostly need caching, rate limits, and a kill switch
- Cloudflare AI Gateway free tier first; add unified billing (+5%) only if the single invoice is worth it.
- If the gateway is your AI governance and security enforcement point (regulated enterprise, agent fleets)
- Portkey under Palo Alto — or self-host its now-open-source gateway if the acquisition roadmap worries you; Kong if you're already a Kong shop.
This analysis is drawn from the Model Router element dossier — the ranked top 5, the comparison matrix, and the complete field of 20 more tools live there, with every source. Data: rt.json (CC BY 4.0).
Every claim above is dated and sourced from the elems dossiers — 1,421 tools tracked across 58 categories, verified 2026-08-06, including the 276 we found dead, renamed, acquired, or sunsetting. Rankings are editorial, never paid — the charter.
Build your stack in 5 questions →