Compare

GoModel vs Portkey

Portkey is an open-source TypeScript gateway plus a hosted platform for logs, budgets, and prompts. Palo Alto Networks acquired it in May 2026 and sells the enterprise edition as Prisma AIRS AI Gateway. In GoModel, the budgets, logs, and dashboard are part of the open-source binary.

Get started with GoModel See the benchmark

4.3×

faster responses

2.2 ms vs 9.6 ms

4.8×

more throughput

4,498 req/s vs 934 req/s

5.6×

faster first token

5.0 ms vs 27.8 ms

2.8×

faster cold start

0.8 s vs 2.3 s

The benchmark

A gateway sits on every request, so its overhead adds up. Same machine, same mock backend, one gateway at a time.

Median latency

p50, chat completions · lower is better

4.3× faster

GoModel2.2 ms

Portkey9.6 ms

Tail latency

p99, chat completions · lower is better

4.2× lower

GoModel6.9 ms

Portkey29.1 ms

Time to first token

p50, streaming chat · lower is better

5.6× faster

GoModel5.0 ms

Portkey27.8 ms

Throughput

peak sustained · higher is better

4.8× more

GoModel4,498 req/s

Portkey934 req/s

Memory

peak RAM · lower is better

1.7× less

GoModel74 MB

Portkey127 MB

Cold start

container start to first request · lower is better

2.8× faster

GoModel0.8 s

Portkey2.3 s

Docker image

compressed pull size · lower is better

2.8× smaller

GoModel21 MB

Portkey58 MB

GoModel 0.1.98 vs Portkey 1.15.2 on an AWS c7i.large (2 vCPU, 4 GiB), run 20260928-043817 (28 September 2026). 20,000 requests per workload at concurrency 10, five randomized trials. API workloads served: GoModel 6/6, Portkey 4/6. Portkey serves /v1/messages only for Anthropic-family providers, so both Anthropic Messages workloads against the benchmark's OpenAI-compatible backend failed. Rerun it with one command.

Beyond speed

Portkey's open-source gateway routes; most of the product lives in its hosted control plane. GoModel ships those parts self-hosted.

	GoModel	Portkey
Self-hosted budgets, usage, logs UI	Free, MIT	Portkey Cloud or Enterprise
Open-source releases	14 in September 2026	Last: v1.15.2, 12 January 2026
Anthropic Messages API	Any provider	Anthropic, Bedrock, Vertex, Azure AI only
Response caching	Exact and semantic, self-hosted	In-memory exact in OSS; semantic hosted
MCP gateway	Free, MIT	Hosted or Enterprise
Providers	34	70+, 1,600+ models
Prompt management	Not included	Versioned templates in Portkey Cloud
SSO and org admin	OIDC SSO in Pro	SSO, SCIM, RBAC on Enterprise

Which one fits?

An honest split.

Pick GoModel if

You need budgets, usage tracking, and a logs dashboard self-hosted, with no enterprise contract.
Claude Code or Anthropic SDK clients must reach any provider over /v1/messages.
You run the gateway yourself and want active releases.
You want one static Go binary with no Node.js runtime to patch.

Pick Portkey if

You need long-tail providers or modalities beyond GoModel's list.
You want logs, prompts, and semantic caching run for you as a service.
You need SAML, SCIM, and org RBAC, or already buy Prisma AIRS from Palo Alto Networks.

Portkey facts are from its repository, docs, and pricing as of October 2026. A Gateway 2.0 branch exists but has not been released.

Try it against your own workload

Benchmarks are an argument; your traffic is the proof. GoModel speaks the OpenAI and Anthropic APIs natively, so pointing an existing app at it is a base-URL change.

1 · Run GoModel

docker run --rm -p 8080:8080 \
  -e LOG_FORMAT=text \
  -e OPENAI_API_KEY="your-openai-key" \
  enterpilot/gomodel

2 · Point your SDK at it

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8080/v1",
    api_key="your-gomodel-key",
)

Full setup, providers, and configuration live in the documentation. Questions? Ask on Discord or book a 30-minute call.