Providers
Four providers. Multiple model-level routes.
Provider support is deliberately selective. Waymesh keeps a curated catalog of model routes whose cost, latency, capability and trust differences give the router a real decision to make. The catalog is smaller than each vendor’s full model list by design.
At a glance
| Provider | Execution | Active catalog models | Private tier | Status |
|---|---|---|---|---|
| OpenAI | Hosted API | gpt-4o-minidefaultgpt-5.6-lunagpt-5.6-terra | Gated out | Implemented |
| Anthropic | Hosted API | claude-opus-4-8defaultclaude-sonnet-5 | Gated out | Implemented |
| Groq | Hosted API — open models | openai/gpt-oss-120bdefaultopenai/gpt-oss-20b | Gated out | Implemented |
| Private / self-hosted | Self-hosted — Ollama | llama3.2default | Eligible | Implemented |
Waymesh routes at the model level inside supported providers. The models shown above are the active curated routes, not every model exposed by each vendor. Existing environment model variables remain compatibility fallbacks for pre-catalog execution; they do not override a model selected by the routing catalog.
Provider detail
OpenAI
openaiHosted API
Catalog routes
- Use case
- A broad commercial model family with several distinct cost and capability points. Waymesh can now choose between multiple OpenAI routes instead of treating the provider as a single fixed model.
- Privacy characteristics
- Third-party execution. Requests leave your infrastructure, so workloads on the private tier are gated out of this route before scoring.
- Availability requirements
- OPENAI_API_KEY
- Reachable api.openai.com
- Catalog route is active and eligible
- Integration status
- Implemented — multi-model catalog routing, live pricing, health check
- Curated Waymesh catalog, not every model OpenAI offers
- GPT-5.6 Luna and Terra are capped at a 272k context window until tier-aware long-context pricing is represented
- Text analysis task type only
Anthropic
anthropicHosted API
Catalog routes
- Use case
- An independent frontier route with more than one model choice. A capability or availability problem at one vendor does not remove the whole top tier, and model-level routing adds another decision inside the provider itself.
- Privacy characteristics
- Third-party execution. Same gating as any hosted API: ineligible for workloads whose policy forbids external egress.
- Availability requirements
- ANTHROPIC_API_KEY
- Reachable api.anthropic.com
- Catalog route is active and eligible
- Integration status
- Implemented — multi-model catalog routing, live pricing, health check
- Curated Waymesh catalog, not every Anthropic model
- Text analysis task type only
Groq
groqHosted API — open models
Catalog routes
- Use case
- A high-speed inference option with two GPT-OSS sizes in the active Waymesh catalog. The smaller route gives the router another lower-cost choice while the 120B route remains the provider default.
- Privacy characteristics
- Third-party execution. Ineligible for private-tier workloads on the same basis as the other hosted routes.
- Availability requirements
- GROQ_API_KEY
- Reachable api.groq.com
- Catalog route is active and eligible
- Integration status
- Implemented — multi-model catalog routing, live pricing, health check
- Curated Waymesh catalog currently includes GPT-OSS 20B and 120B
- Context ceilings differ from the frontier routes and are enforced as eligibility constraints
- Text analysis task type only
Private / self-hosted
ollamaSelf-hosted — Ollama
Catalog routes
- Use case
- Execution on infrastructure you control, so a workload whose policy forbids third-party egress has a legal route rather than no route.
- Privacy characteristics
- Prompts stay on your endpoint. The adapter refuses a remote plaintext HTTP endpoint outright — sending prompts in the clear would contradict the private tier it exists to serve — unless that refusal is explicitly overridden.
- Availability requirements
- OLLAMA_BASE_URL reachable from the worker
- Local, private-network or HTTPS endpoint
- Catalog route is active and eligible
- Integration status
- Implemented — real adapter, endpoint safety classification, health check
- One curated Ollama model is currently declared
- Throughput and latency are whatever your hardware provides
- Text analysis task type only
Roadmap
The catalog grows when a new route changes the decision.
New models and providers are added when a real workload needs a routing alternative the current set cannot offer — a lower cost point, a context ceiling, a latency floor, a modality or a residency requirement. That is a better reason than coverage for its own sake.
A route appears on this page after its integration exists and the model has been verified for the deployment path. Waymesh does not present a vendor’s entire upstream catalog as supported simply because the API can name it.
Waymesh is independent of every provider it routes to and is not affiliated with, endorsed by or partnered with any of them. Provider names appear here to describe compatibility only.
Tell us which route your workload needsWhich route is right for a given workload is a measurement, not an opinion. Shadow mode lets Waymesh evaluate those choices against your own traffic before routing policy changes.
Apply for Design Partner Access