OpenAI Model Catalog

Werker Health proxies every generally available OpenAI model. Use the exact model string—no prefixing required—after pointing the SDK at https://api.werker.health/v1.

✅ Newly announced models (including GPT-5 variants) become available on Werker Health within minutes of OpenAI launch. See the OpenAI Models reference for the authoritative list.

Featured GPT-5 Lineup

Model	Recommended Use
gpt-5	Frontier flagship for complex coding, agentic reasoning, and cross-domain problem solving.
gpt-5-mini	Cost-efficient GPT-5 tier for high-volume tasks with well-defined outputs.
gpt-5-nano	Fastest GPT-5 deployment for latency-sensitive automations.
gpt-5-pro	Premium variant tuned for more precise responses and deeper reasoning.

GPT-4 Series

Model	Recommended Use
gpt-4.1	Production-grade reasoning with strong multimodal performance.
gpt-4.1-mini	Balanced quality and price for copilots and summarization.
gpt-4o	Multimodal flagship optimized for cost, latency, and speech.
gpt-4o-mini	Scaled deployments needing strong quality under tight latency budgets.
gpt-4o-realtime-preview	Audio-first experiences and real-time assistants.

o-Series (High Planning)

Model	Recommended Use
o4-mini	Long-context, tool-heavy agents benefiting from explicit reasoning traces.
o3-mini	Lightweight planning, evaluation, and oversight workflows.

Specialized & Legacy Models

Model	Recommended Use
gpt-4o-audio-preview	Speech-centric assistants that transcribe, reason, and respond vocally.
gpt-4o-mini-transcribe	Medical and claims transcription with high accuracy.
gpt-4o-code	Code generation, reviews, and refactoring tasks.
gpt-3.5-turbo	Legacy workloads not yet migrated to GPT-4o/5.
text-embedding-3-large / text-embedding-3-small	Semantic search, retrieval, and analytics pipelines.

Managing Model Access

• Configure per-environment allowlists to control which model families can receive PHI.
• Use Werker’s billing dashboards to monitor cost per model and set budgets.
• Combine metadata fields with Werker audit logs to track which application feature invoked each model.

Rolling Updates

• Automatic availability: Newly released OpenAI models appear without app changes.
• Version pinning: Set explicit model strings in production to avoid surprises; use staging to evaluate new versions.
• Testing new models: Clone workloads in Werker’s sandbox environment and compare output quality before promoting to production.

Sample Usage

Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.werker.health/v1",
    api_key="WERKER_HEALTH_API_KEY",
)

plan = client.responses.create(
    model="gpt-5",
    input=[
        {"role": "system", "content": "You generate discharge plans for clinicians."},
        {"role": "user", "content": "Create discharge instructions for a mild asthma flare."},
    ],
)

print(plan.output_text)

TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.werker.health/v1",
  apiKey: "WERKER_HEALTH_API_KEY",
});

const plan = await client.responses.create({
  model: "gpt-5",
  input: [
    { role: "system", content: "You generate discharge plans for clinicians." },
    { role: "user", content: "Create discharge instructions for a mild asthma flare." },
  ],
});

console.log(plan.output_text);