{
 "sym": "Rt",
 "updated": "2026-08-06",
 "verdict": "OpenRouter, and it isn't close for the default case \u2014 400+ models behind one key, 25 trillion tokens a week (5x in six months), 8M developers, and a $113M Series B at $1.3B led by CapitalG (May 2026) that made it the category's only unicorn. Pay its ~5.5% credit fee as rent for never touching provider plumbing. LiteLLM is the pick when keys and traffic must stay in your VPC \u2014 the open-source gateway standard with a 2026 Rust core. Vercel AI Gateway wins on price (genuinely 0% markup, free BYOK) if you live in the AI SDK; Cloudflare AI Gateway when you're already on Cloudflare and want the free edge control plane; Portkey when enterprise governance is the job \u2014 bought by Palo Alto Networks in May 2026, which is both its endorsement and its risk.",
 "top5": [
  {
   "rank": 1,
   "name": "OpenRouter",
   "maker": "OpenRouter, Inc.",
   "url": "https://openrouter.ai",
   "docs": "https://openrouter.ai/docs",
   "pricing": "Pass-through token pricing (0% markup) \u00b7 5.5% fee on credit purchases ($0.80 min; 5% via crypto) \u00b7 BYOK free to 1M req/mo, then 5%",
   "best_for": "Any team that wants every model behind one key and one invoice today, with zero infrastructure to run.",
   "why": "The category's runaway aggregator: 25T tokens/week (up 5x in six months, on pace for a quadrillion/year), 8M+ developers, 400+ models across 60+ providers, and a $113M Series B at ~$1.3B (May 2026) with CapitalG, NVIDIA, ServiceNow, MongoDB, Snowflake and Databricks all on the round. Edge routing adds ~25ms; pricing is pure pass-through, so the only cost is the credit fee. Sacra pegs annualized platform revenue at ~$50M by Mar 2026 \u2014 proof the ~5% take rate model works at scale.",
   "watch": "That 5.5% compounds painfully at volume \u2014 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.",
   "evidence": [
    {
     "stat": "$113M Series B led by CapitalG at ~$1.3B; 25T tokens/week, 8M devs, 400+ models (May 26\u201328, 2026)",
     "src": "https://openrouter.ai/blog/announcements/series-b/"
    },
    {
     "stat": "Valuation 2.4x in a year ($547M Jun 2025 \u2192 $1.3B May 2026); 100T tokens/month (TechCrunch, May 26, 2026)",
     "src": "https://techcrunch.com/2026/05/26/openrouter-more-than-doubles-valuation-to-1-3b-in-a-year/"
    },
    {
     "stat": "~$50M annualized revenue by Mar 2026 on ~5% take; ~25ms added edge latency (Sacra)",
     "src": "https://sacra.com/c/openrouter/"
    }
   ],
   "tile_note": "one api, every model"
  },
  {
   "rank": 2,
   "name": "LiteLLM",
   "maker": "BerriAI (YC W23; MIT OSS)",
   "url": "https://www.litellm.ai",
   "docs": "https://docs.litellm.ai",
   "pricing": "Open source free forever (self-host) \u00b7 Enterprise custom annual, sized to request capacity \u2014 never per-token \u00b7 30-day trial",
   "best_for": "Teams that need the router inside their own VPC \u2014 keys, logs, and spend controls that never leave your infrastructure.",
   "why": "The open-source gateway standard: 53.8k GitHub stars, 1,000+ contributors, 240M+ Docker pulls, and testimonials from Netflix, NVIDIA, Okta and Stripe. Claims 140+ providers and ~1,900 models behind one OpenAI-compatible API with virtual keys, team budgets, fallbacks and Prometheus metrics all in the free tier. The 2026 Rust-core gateway answered the performance critics: 0.66ms added at p99 (vendor benchmark, ~2,800 RPS).",
   "watch": "You run it \u2014 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.",
   "evidence": [
    {
     "stat": "53.8k stars, 9.8k forks, 40k+ commits; 100+ provider endpoints (GitHub, Aug 2026)",
     "src": "https://github.com/BerriAI/litellm"
    },
    {
     "stat": "240M+ Docker pulls, 1,005+ contributors; 140+ providers, 1,892 models claimed (vendor, Aug 2026)",
     "src": "https://www.litellm.ai/"
    },
    {
     "stat": "Rust gateway adds 0.66ms p99 at ~2,800 RPS \u2014 vendor benchmark, Aug 2026",
     "src": "https://www.litellm.ai/"
    }
   ],
   "tile_note": "open-source gateway standard"
  },
  {
   "rank": 3,
   "name": "Vercel AI Gateway",
   "maker": "Vercel",
   "url": "https://vercel.com/ai-gateway",
   "docs": "https://vercel.com/docs/ai-gateway",
   "pricing": "0% markup, 0% platform fee \u2014 provider list price via credits \u00b7 BYOK free \u00b7 free tier (subset of models) \u00b7 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.",
   "why": "The price disruptor: GA since Aug 21, 2025 with sub-20ms latency, hundreds of models, automatic failover \u2014 and genuinely zero markup, including on BYOK, where OpenRouter charges 5%+. Its monthly production index (a real telemetry set: open-weight models hit 29% of gateway tokens on <4% of spend, Jul 2026) shows serious traffic already flows through it. Vercel subsidizes the gateway to sell the surrounding cloud, which is exactly why it's cheap.",
   "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.",
   "evidence": [
    {
     "stat": "GA Aug 21, 2025: zero markup, 100+ models, sub-20ms latency",
     "src": "https://vercel.com/blog/ai-gateway-is-now-generally-available"
    },
    {
     "stat": "\"AI Gateway charges no markup and no platform fee on tokens\"; BYOK with no fee (docs, updated Aug 1, 2026)",
     "src": "https://vercel.com/docs/ai-gateway/pricing"
    },
    {
     "stat": "Production index Jul 2026: token volume +29% MoM; open-weight models 29% of tokens, <4% of spend",
     "src": "https://vercel.com/blog/ai-gateway-production-index-july-2026"
    }
   ],
   "tile_note": "zero-markup managed gateway"
  },
  {
   "rank": 4,
   "name": "Cloudflare AI Gateway",
   "maker": "Cloudflare",
   "url": "https://www.cloudflare.com/products/ai-gateway/",
   "docs": "https://developers.cloudflare.com/ai-gateway/",
   "pricing": "Core features free (analytics, caching, rate limits, dynamic routing) \u00b7 unified billing +5% on credit purchases \u00b7 Workers Paid ($5/mo) lifts log caps to 10M/gateway",
   "best_for": "Teams already on Cloudflare who want a free, edge-speed control plane \u2014 caching, rate limiting, DLP \u2014 in front of any model API.",
   "why": "The strongest free tier in the category, run on the network a fifth of the web already transits. The Aug 27, 2025 refresh added dynamic routing (A/B tests, per-user limits, model chaining), Secrets Store key management, DLP scanning in the AI firewall, and 350+ models across major providers \u2014 with optional unified billing at a 5% credit fee only if you want one invoice.",
   "watch": "Routing intelligence is thinner than dedicated routers \u2014 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.",
   "evidence": [
    {
     "stat": "Core AI Gateway features free; unified billing 5% surcharge on credits; DLP free on all plans (docs, 2026)",
     "src": "https://developers.cloudflare.com/ai-gateway/reference/pricing/"
    },
    {
     "stat": "Aug 27, 2025 refresh: dynamic routing, unified billing beta, Secrets Store, 350+ models across providers",
     "src": "https://blog.cloudflare.com/ai-gateway-aug-2025-refresh"
    }
   ],
   "tile_note": "free edge control plane"
  },
  {
   "rank": 5,
   "name": "Portkey",
   "maker": "Portkey \u2192 Palo Alto Networks (closed May 29, 2026)",
   "url": "https://portkey.ai",
   "docs": "https://portkey.ai/docs",
   "pricing": "Gateway fully open source (Mar 2026) \u00b7 Developer free (10k logs/mo) \u00b7 Production $49/mo (100k logs, +$9/100k) \u00b7 Enterprise custom (VPC, SOC2/HIPAA)",
   "best_for": "Enterprises that treat the gateway as a governance layer \u2014 guardrails, RBAC, budgets, prompt management, audit \u2014 not just a switchboard.",
   "why": "The governance-first gateway at real scale: 1T+ tokens and 120M+ requests daily across 24,000+ organizations, $180M+ in managed AI spend (Mar 2026), with the full enterprise gateway open-sourced that same month. Palo Alto Networks acquired it (closed May 29, 2026) to anchor Prisma AIRS \u2014 the clearest possible signal that routing is becoming the enforcement point for AI security.",
   "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.",
   "evidence": [
    {
     "stat": "1T+ tokens/day, 120M+ requests/day, 24,000+ orgs; gateway fully open-sourced (Mar 24, 2026)",
     "src": "https://www.globenewswire.com/news-release/2026/03/24/3261574/0/en/portkey-s-gateway-is-now-fully-open-source-processing-over-1-trillion-tokens-every-day.html"
    },
    {
     "stat": "Palo Alto Networks completed acquisition May 29, 2026; Portkey becomes core of Prisma AIRS (terms undisclosed)",
     "src": "https://www.paloaltonetworks.com/company/press/2026/palo-alto-networks-completes-acquisition-of-portkey-to-secure-ai-agents"
    },
    {
     "stat": "Production tier $49/mo for 100k logs; enterprise adds VPC deploy, SSO, custom guardrails (pricing page, Aug 2026)",
     "src": "https://portkey.ai/pricing"
    }
   ],
   "tile_note": "governance gateway, now panw"
  }
 ],
 "matrix": {
  "cols": [
   "Take rate / fees",
   "Hosting",
   "Open source",
   "Catalog",
   "Added latency",
   "Routing depth",
   "Enterprise"
  ],
  "rows": [
   [
    "OpenRouter",
    "0% markup \u00b7 5.5% credit fee \u00b7 BYOK 5% >1M req",
    "Managed edge",
    "No",
    "400+ models \u00b7 60+ providers",
    "~25 ms",
    "Fallbacks \u00b7 price/throughput sorts \u00b7 ZDR routing",
    "Workspaces \u00b7 spend mgmt \u00b7 ZDR"
   ],
   [
    "LiteLLM",
    "Free \u00b7 enterprise custom (not per-token)",
    "Self-host (+ cloud)",
    "Yes (MIT)",
    "140+ providers \u00b7 ~1,900 models (claimed)",
    "0.66 ms p99 (vendor)",
    "Fallbacks \u00b7 load-balance \u00b7 budgets \u00b7 caching",
    "SSO \u00b7 SCIM \u00b7 audit \u00b7 air-gap"
   ],
   [
    "Vercel AI Gateway",
    "0% markup \u00b7 0% BYOK \u00b7 metered add-ons",
    "Managed CDN",
    "No",
    "100s of models",
    "sub-20 ms",
    "Failover \u00b7 provider ordering \u00b7 per-request ZDR",
    "Invoiced billing \u00b7 allowlists"
   ],
   [
    "Cloudflare AI Gateway",
    "Core free \u00b7 +5% unified billing",
    "Managed edge",
    "No",
    "350+ models",
    "edge (unpublished)",
    "Dynamic routing \u00b7 caching \u00b7 rate limits",
    "DLP \u00b7 Secrets Store \u00b7 Logpush"
   ],
   [
    "Portkey",
    "OSS free \u00b7 $49/mo prod \u00b7 ent custom",
    "Both",
    "Yes (gateway)",
    "All major providers",
    "low (edge deploys)",
    "Guardrails \u00b7 load-balance \u00b7 canary \u00b7 prompt mgmt",
    "RBAC \u00b7 SSO \u00b7 VPC \u00b7 SOC2/HIPAA"
   ],
   [
    "Requesty",
    "5% all-in markup",
    "Managed",
    "Partial",
    "600+ models",
    "<14 ms failover",
    "Smart routing \u00b7 semantic cache \u00b7 guardrails",
    "SSO \u00b7 spend controls (EU angle)"
   ],
   [
    "Helicone AI Gateway",
    "OSS free \u00b7 cloud 0% markup (beta)",
    "Both",
    "Yes (Rust)",
    "100+ models",
    "light (Rust, vendor)",
    "Cheapest-provider routing \u00b7 fallbacks \u00b7 caching",
    "Observability-native \u00b7 SLA on cloud"
   ],
   [
    "Bifrost",
    "OSS free \u00b7 enterprise custom",
    "Self-host",
    "Yes (Apache-2.0)",
    "1,000+ models \u00b7 23+ providers",
    "<15 \u00b5s @5k RPS (vendor)",
    "Adaptive load-balance \u00b7 semantic cache \u00b7 MCP",
    "Cluster mode \u00b7 governance \u00b7 budgets"
   ]
  ]
 },
 "rules": [
  {
   "if": "You just hit two models in production and have no infra team",
   "then": "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 \u2014 or procurement says self-host",
   "then": "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",
   "then": "Vercel AI Gateway \u2014 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",
   "then": "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)",
   "then": "Portkey under Palo Alto \u2014 or self-host its now-open-source gateway if the acquisition roadmap worries you; Kong if you're already a Kong shop."
  }
 ],
 "field": [
  {
   "name": "Requesty",
   "maker": "Requesty (Amsterdam)",
   "note": "EU-flavored OpenRouter alternative: 600+ models, 5% all-in markup, <14ms failover, 225B+ tokens/day claimed; $3M seed led by 20VC (Sep 2025)",
   "url": "https://www.requesty.ai",
   "oss": false,
   "entry": "5% markup",
   "status": "active"
  },
  {
   "name": "Helicone AI Gateway",
   "maker": "Helicone (YC W23)",
   "note": "Open-source Rust gateway from the observability company; cloud version with 0%-markup passthrough billing launched to beta Sep 2025",
   "url": "https://github.com/Helicone/ai-gateway",
   "oss": true,
   "entry": "free + tokens",
   "status": "active"
  },
  {
   "name": "Bifrost",
   "maker": "Maxim AI",
   "note": "Go gateway marketing itself on speed \u2014 '50x faster than LiteLLM', <15\u00b5s overhead at 5k RPS (vendor benchmark), 1,000+ models, MCP built in",
   "url": "https://github.com/maximhq/bifrost",
   "oss": true,
   "entry": "free + tokens",
   "status": "active"
  },
  {
   "name": "Kong AI Gateway",
   "maker": "Kong Inc.",
   "note": "The API-gateway incumbent's AI layer: semantic routing/caching, prompt firewalling, multi-LLM plugins on Kong Gateway/Konnect",
   "url": "https://konghq.com/products/kong-ai-gateway",
   "oss": true,
   "entry": "OSS free \u00b7 Konnect tiers",
   "status": "active"
  },
  {
   "name": "Microsoft Foundry Model Router",
   "maker": "Microsoft",
   "note": "Hyperscaler-native learned router, GA Nov 2025 (version 2025-11-18): picks among 28 models incl. GPT-5.x, Claude, DeepSeek, Grok inside Azure",
   "url": "https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/model-router",
   "oss": false,
   "entry": "Azure usage",
   "status": "active"
  },
  {
   "name": "Bedrock Intelligent Prompt Routing",
   "maker": "AWS",
   "note": "Per-prompt routing between models of a family inside Bedrock; AWS-only answer to the router question",
   "url": "https://aws.amazon.com/bedrock/",
   "oss": false,
   "entry": "AWS usage",
   "status": "active"
  },
  {
   "name": "LLM Gateway",
   "maker": "LLMGateway (OSS)",
   "note": "Open-source OpenRouter alternative; 0% fee with own keys, 5% on credits \u2014 publishes the category's best fee-comparison research",
   "url": "https://llmgateway.io",
   "oss": true,
   "entry": "free + tokens",
   "status": "active"
  },
  {
   "name": "TrueFoundry AI Gateway",
   "maker": "TrueFoundry",
   "note": "Enterprise self-hosted gateway (K8s-native); prolific comparison-content marketer aiming at LiteLLM/Portkey buyers",
   "url": "https://www.truefoundry.com",
   "oss": false,
   "entry": "custom",
   "status": "active"
  },
  {
   "name": "Eden AI",
   "maker": "Eden AI",
   "note": "Unified API beyond LLMs (vision, OCR, speech); 5.5% fee on credits \u2014 broader but shallower than LLM-first routers",
   "url": "https://www.edenai.co",
   "oss": false,
   "entry": "credits + 5.5%",
   "status": "active"
  },
  {
   "name": "Not Diamond",
   "maker": "Not Diamond",
   "note": "Trained per-prompt quality router (pick the best model, not just the cheapest); shipped a coding-agent router in 2026 \u2014 niche but alive",
   "url": "https://www.notdiamond.ai",
   "oss": false,
   "entry": "free tier + usage",
   "status": "active"
  },
  {
   "name": "Arch",
   "maker": "Katanemo",
   "note": "Open-source agent-native proxy with its own Arch-Router model for preference-based routing",
   "url": "https://github.com/katanemo/archgw",
   "oss": true,
   "entry": "free + tokens",
   "status": "active"
  },
  {
   "name": "Apache APISIX AI Gateway",
   "maker": "Apache Software Foundation",
   "note": "AI plugins (ai-proxy, rate-limiting, prompt guard) on the APISIX API gateway; solid if APISIX is already your edge",
   "url": "https://apisix.apache.org",
   "oss": true,
   "entry": "free",
   "status": "active"
  },
  {
   "name": "Higress",
   "maker": "Alibaba (OSS)",
   "note": "Envoy-based cloud-native AI gateway, big in China deployments; token-aware routing and model failover",
   "url": "https://higress.io",
   "oss": true,
   "entry": "free",
   "status": "active"
  },
  {
   "name": "new-api",
   "maker": "QuantumNous (OSS)",
   "note": "Popular Chinese-community multi-provider aggregator/reseller panel in the One API lineage",
   "url": "https://github.com/QuantumNous/new-api",
   "oss": true,
   "entry": "free + tokens",
   "status": "active"
  },
  {
   "name": "RouteLLM",
   "maker": "LMSYS (research OSS)",
   "note": "The academic cost-quality router that popularized learned routing (2024); repo largely quiet since \u2014 ideas absorbed by commercial routers",
   "url": "https://github.com/lm-sys/RouteLLM",
   "oss": true,
   "entry": "free",
   "status": "fading"
  },
  {
   "name": "Glama Gateway",
   "maker": "Glama",
   "note": "Low-fee gateway attached to an MCP-centric workspace; small but liked by indie agent builders",
   "url": "https://glama.ai/gateway",
   "oss": false,
   "entry": "usage-based",
   "status": "active"
  },
  {
   "name": "Martian",
   "maker": "Martian (withmartian)",
   "note": "Invented the 'model router' pitch ($9M, NEA/Prosus 2023; Accenture 2024) \u2014 has pivoted to AI interpretability research; router no longer the product",
   "url": "https://withmartian.com",
   "oss": false,
   "entry": "\u2014",
   "status": "fading"
  },
  {
   "name": "Unify",
   "maker": "Unify AI",
   "note": "Benchmark-driven LLM router (a16z-backed, 2024) \u2014 router retired; company pivoted to 'AI teammates' workflow agents by 2026",
   "url": "https://unify.ai",
   "oss": false,
   "entry": "\u2014",
   "status": "dead"
  },
  {
   "name": "TensorZero",
   "maker": "TensorZero",
   "note": "Rust gateway + LLMOps stack; archived its repo and shut down Jun 12, 2026, returning remaining capital \u2014 the category's starkest OSS casualty",
   "url": "https://github.com/tensorzero/tensorzero",
   "oss": true,
   "entry": "\u2014",
   "status": "dead"
  },
  {
   "name": "MLflow AI Gateway",
   "maker": "Databricks (OSS)",
   "note": "Deployments/gateway module inside MLflow; maintained but overshadowed \u2014 Databricks invests via Mosaic AI Gateway instead",
   "url": "https://mlflow.org",
   "oss": true,
   "entry": "free",
   "status": "fading"
  }
 ],
 "signals": [
  {
   "fact": "OpenRouter raised a $113M Series B at ~$1.3B led by CapitalG (May 26\u201328, 2026) \u2014 the category's first unicorn; weekly tokens 5T \u2192 25T in six months",
   "src": "https://openrouter.ai/blog/announcements/series-b/"
  },
  {
   "fact": "Consolidation wave: Palo Alto Networks acquired Portkey (closed May 29, 2026) into Prisma AIRS; TensorZero shut down Jun 12, 2026; Martian and Unify both pivoted away from routing",
   "src": "https://www.paloaltonetworks.com/company/press/2026/palo-alto-networks-completes-acquisition-of-portkey-to-secure-ai-agents"
  },
  {
   "fact": "Take rates converged: 0% token markup is now table stakes; money is made on ~5\u20135.5% credit/platform fees (OpenRouter 5.5%, Cloudflare 5%, Eden 5.5%) \u2014 and Vercel's 0%/0% undercuts them all (fee survey updated Aug 5, 2026)",
   "src": "https://llmgateway.io/blog/ai-gateway-fees-compared"
  },
  {
   "fact": "Hyperscalers ship routing natively: Microsoft Foundry Model Router GA Nov 2025 now routes across 28 models including Claude, DeepSeek and Grok \u2014 squeezing standalone routers from above",
   "src": "https://learn.microsoft.com/en-us/azure/foundry/foundry-models/whats-new-model-router"
  },
  {
   "fact": "Routers are the arbitrage layer: open-weight models hit 29% of Vercel AI Gateway tokens on under 4% of spend; Anthropic took 61% of spend on 32% of tokens (Jul 2026 production index)",
   "src": "https://vercel.com/blog/ai-gateway-production-index-july-2026"
  },
  {
   "fact": "Portkey open-sourced its full enterprise gateway (Mar 24, 2026) at 1T+ tokens/day across 24,000+ orgs \u2014 governance features are commoditizing fast",
   "src": "https://www.globenewswire.com/news-release/2026/03/24/3261574/0/en/portkey-s-gateway-is-now-fully-open-source-processing-over-1-trillion-tokens-every-day.html"
  }
 ],
 "notes": "Ranking criteria: effective take rate at real volume, catalog breadth, added latency, governance depth, and survivability \u2014 this category had a brutal 2026 (one shutdown, one acquisition, two pivots), so 'will it exist next year' is a first-order criterion. Conflicts resolved: (1) OpenRouter's ~$50M annualized revenue is Sacra's estimate, not company-disclosed \u2014 treated as directional; (2) a Medium post claiming Martian 'nearing $1.3B valuation' is unsourced and contradicted by its pivot to interpretability research \u2014 rejected; (3) Bifrost's '40x/50x faster than LiteLLM' and LiteLLM's '0.66ms p99' are both vendor benchmarks predating each other's latest rewrites \u2014 reported as claims, not facts; (4) model-count claims (400+, 600+, 1,892, 1,000+) are vendor-reported and drift weekly. Catalog counts and uptime figures are single-source (vendor pages). Adjacent elements: inference clouds serving open weights (Together, Fireworks, Groq, DeepInfra) are providers routers route TO \u2192 Ow; GPT-5-style in-product auto-routing lives inside the model app \u2192 Fm; observability-first tools (Langfuse, Helicone-as-monitor, Braintrust) \u2192 Ev; prompt/response filtering as a product \u2192 Gd. Cloud-desk aggregators reselling coding-agent subscriptions are out of scope (\u2192 Ca).",
 "sources": [
  "https://openrouter.ai/blog/announcements/series-b/",
  "https://techcrunch.com/2026/05/26/openrouter-more-than-doubles-valuation-to-1-3b-in-a-year/",
  "https://sacra.com/c/openrouter/",
  "https://openrouter.ai/docs/faq",
  "https://www.litellm.ai/",
  "https://www.litellm.ai/pricing",
  "https://github.com/BerriAI/litellm",
  "https://vercel.com/blog/ai-gateway-is-now-generally-available",
  "https://vercel.com/docs/ai-gateway/pricing",
  "https://vercel.com/blog/ai-gateway-production-index-july-2026",
  "https://developers.cloudflare.com/ai-gateway/reference/pricing/",
  "https://blog.cloudflare.com/ai-gateway-aug-2025-refresh",
  "https://portkey.ai/pricing",
  "https://www.paloaltonetworks.com/company/press/2026/palo-alto-networks-completes-acquisition-of-portkey-to-secure-ai-agents",
  "https://www.globenewswire.com/news-release/2026/03/24/3261574/0/en/portkey-s-gateway-is-now-fully-open-source-processing-over-1-trillion-tokens-every-day.html",
  "https://www.requesty.ai/",
  "https://www.requesty.ai/blog/requesty-raises-3m",
  "https://llmgateway.io/blog/ai-gateway-fees-compared",
  "https://github.com/maximhq/bifrost",
  "https://learn.microsoft.com/en-us/azure/foundry/foundry-models/whats-new-model-router",
  "https://byteiota.com/tensorzero-shuts-down-what-oss-llmops-cant-survive/",
  "https://www.helicone.ai/blog/ptb-gateway-launch",
  "https://unify.ai",
  "https://withmartian.com"
 ],
 "element": {
  "number": 4,
  "name": "Model Router",
  "group": "Intelligence",
  "essential": false,
  "edition": "v2026.Q3",
  "revision": "r7",
  "license": "CC BY 4.0 \u2014 cite elems.ai",
  "url": "https://elems.ai/e/rt.html"
 }
}