ai-gateway on tomrochette.com

OpenCode Zen

tom@tomrochette.com (Tom Rochette) — Sat, 26 Sep 2026 00:00:00 +0000

OpenCode Zen is the OpenCode (Anomaly) team’s curated pay-per-use AI gateway, one API key over a tested catalog that spans open models and the Claude, GPT, Gemini, and Grok lines.

What it is
#

Zen is a billing and routing gateway, not a model: you add credits, get a key, and call OpenAI-compatible, Anthropic-compatible, Google, or Responses endpoints under opencode.ai/zen/v1. The team benchmarks each model/provider pair and serves what passes, their answer to getting a degraded version of a model through a generic router. It also sells to teams (workspaces, roles, member spending caps, model toggles, BYOK for OpenAI and Anthropic keys), and free stealth and promo models rotate through the catalog.

Status
#

Active, with heavy catalog churn managed through a public deprecation table (docs updated 2026-09-25), and workspaces free during the beta with team pricing unannounced. The Reddit thread “Opencode Zen is astoundingly more expensive than OpenRouter” (July 2026) is the sharpest public criticism and remains the note’s key stress test. Zen sells curation and reliability, and community measurements say that premium can exceed 4x the cheapest gateway on identical models.

Strengths
#

Every endpoint is benchmarked for coding-agent use, so you avoid badly served providers.
A stable of free models (Big Pickle, Space Bunny Free, MiMo Free, Nemotron Free) with documented data caveats.
Markups are claimed to be zero beyond pass-through processing fees (4.4% + $0.30 per card transaction).
US-hosted with zero-retention on most models, exceptions listed per model.
Same balance backs the Go subscription’s overflow, and the key works from any agent.

Cautions
#

A July 2026 community test measured Zen at more than 4x OpenRouter on Kimi K2.6 and MiniMax M2.7.
The gap persists per model as of 2026-09-26: Zen lists GLM-5.3-Flash at $0.15/$0.50 per 1M tokens while OpenRouter’s catalog shows about $0.05/$0.14.
Auto-reload adds $20 whenever your balance drops below $5, which can overrun a monthly budget you set.
Deprecated models vanish on published dates, so configs need migrations, and free stealth models may use your data to improve the model during their free window.

Pricing
#

Pay-as-you-go per 1M tokens, as of 2026-09-26: GLM-5.3 $1.40/$4.40, GLM-5.3-Flash $0.15/$0.50, Kimi K3 $3/$15, Qwen3.7 Plus $0.40/$1.60, DeepSeek V4.1 Flash $0.30/$1.20, MiniMax M3 $0.30/$1.20. Frontier lines: Claude Sonnet 5 $2/$10, Claude Opus 5.5 $4/$20, GPT 5.5 $5/$30, Gemini 3.8 Flash $1.50/$7.50, Grok 4.7 $2/$6. Jev 1.13 charges $0.042 per input with free output, and nine models are free including the Big Pickle and Space Bunny stealth models, all for a limited time. Auto-reload charges $20 when the balance falls below $5, and OpenCode itself uses low-cost models to generate session titles on your bill.

Price history
#

Date	Plan	Change	Source
2026-09-26	Pay-per-use catalog	Baseline: per-model rates as published (GLM-5.3 $1.40/$4.40, Kimi K3 $3/$15, Jev 1.13 $0.042 in)	https://opencode.ai/docs/zen/

Compared to
#

OpenCode Go (../opencode-go/index.md) is the sibling subscription: $10/month for capped open-model usage, no frontier models, same balance; choose it when you want a ceiling. OpenRouter (../openrouter/index.md) is the breadth play: hundreds of models, passthrough pricing, 5.5% credit fee; choose it when price per token on open models matters more than curation. Holding your own provider keys is cheapest for a single provider, and Zen’s BYOK only covers OpenAI and Anthropic.

Bottom line
#

Recommended for OpenCode-centric developers who want vetted endpoints, free models to fall back on, and US-hosted zero-retention. Not for price-sensitive, high-volume runs on open models, where OpenRouter wins on raw per-token cost. My disagreeable claim: the curation premium is worth paying only when a mis-served provider would silently degrade your agent, and most coding work is not that sensitive.

Changes
#

2026-09-26 - Created.

References
#

https://opencode.ai/docs/zen/ - per-model price table, free models, auto-reload, deprecations, privacy (200).
https://docs.docker.com/ai/docker-agent/providers/opencode-zen/ - third-party integration doc, Zen vs Go billing table (200).
https://www.reddit.com/r/opencodeCLI/comments/1syog9v/opencode_zen_is_astoundingly_more_expensive_than/ - critical 4x price claim vs OpenRouter (direct fetch 403; content read via search).
https://costgoat.com/pricing/openrouter/ - OpenRouter per-model prices used for the GLM-5.3-Flash comparison (200, as of 2026-09-25).
https://opencode.ai/docs/go - shared balance and “Use balance” fallback mechanics (200).
https://www.truefoundry.com/blog/openrouter-pricing - OpenRouter fee structure used as the price counterpoint (200).

Requesty

tom@tomrochette.com (Tom Rochette) — Sat, 26 Sep 2026 00:00:00 +0000

Requesty is an EU-hosted managed AI gateway that puts 600+ models behind one OpenAI-compatible endpoint and charges a flat 5% markup on upstream model spend.

The 5% markup is the whole business model: keep your OpenAI SDK code, point the base URL at router.requesty.ai/v1, and pay model cost plus 5% for routing, caching, and governance.

What it is
#

I read it as pass-through billing with a gateway attached: hosted-only, no subscription tiers, no seat fees, no minimum spend. You get routing policies, fallback chains, semantic caching, budget caps, PII detection, and an MCP gateway over one key. Their own comparison post says the gateway is built in Rust, and EU residency in Frankfurt is central to their positioning.

Status
#

Alive and funded: a $3M seed led by 20VC was announced 2025-09-26, with Tapestry VC, Insiders Ventures, and Tiny Supercomputer, and Business Insider covered the pitch deck.
The site footer reads © 2026 Requesty Ltd and their blog shipped multiple gateway comparisons through June and September 2026, yet an hn.algolia.com search for “requesty.ai” returns no story about the company, so their footprint is SEO and docs, not Hacker News.

Strengths
#

Pricing is one line long: a $10/1M token model costs $10.50 through them, and the free tier gives 200 requests/day on free models with no credit card.
EU (Frankfurt) data residency is included on every plan.
Enterprise adds SSO, RBAC, audit logs, guardrails, and service accounts for CI/CD.

Cautions
#

The 5% scales with your spend, so the fee grows exactly when your usage does, unlike flat-priced access.
A competitor-authored guide (TrueFoundry, which sells a rival gateway) notes Requesty is hosted-only with no self-hosting, had SOC 2 Type II still in progress (expected Q3 2026), and keeps self-service prompt logging on for up to 30 days by default, with organization-wide zero retention available only on written request.
Their own properties disagree on catalog size: 600+ on the pricing page versus 400+ in their June 2026 comparison.
The “caching makes us net-negative on cost” claim (40-60% hit rates) is vendor math I could not verify independently.

Pricing
#

Free: $0, free models only, 200 requests/day, routing, caching, spend tracking, and EU residency included.
Pay as you go: flat 5% markup on upstream model cost, all 600+ models, budget caps, and MCP gateway, with BYOK carrying 0% markup per their docs (as of 2026-09-26).
Enterprise: custom, with SSO (Okta, Azure AD, Google Workspace, custom OIDC), RBAC, approved-model policies, and custom SLAs.

Price history
#

Date	Plan	Change	Source
2026-09-26	Pay as you go	Baseline: flat 5% markup on upstream spend, free tier at 200 requests/day	requesty.ai/pricing

Compared to
#

OpenRouter: a 5.5% fee on credit purchases plus a $0.80 minimum and 365-day credit expiry per Requesty’s own comparison; pick OpenRouter for catalog breadth and instant start, Requesty for governance and EU residency.
LiteLLM: a free self-hosted proxy with zero markup; pick it if you have DevOps capacity and want no vendor in the billing path.
Synthetic (Synthetic): a flat subscription for open-weight coding models, not a general gateway; pick it for agent coding at a fixed monthly cost.

Bottom line
#

Recommended for teams routing production multi-provider traffic who want EU residency, budget governance, and a bill finance can read as a percentage. Not for solo developers who only want cheap model access, or anyone who needs self-hosting. My most contestable claim: without repetitive traffic that actually hits the cache, you are just paying 5% for a proxy.

Changes
#

2026-09-26 - Created.

References
#

https://requesty.ai/pricing - 5% markup quote, tier table, 200 requests/day free tier, EU residency (as of 2026-09-26)
https://www.requesty.ai/blog/requesty-raises-3m - $3M seed led by 20VC, investor list, EU positioning (published 2025-09-26)
https://www.requesty.ai/blog/best-llm-routing-platforms-compared-2026-requesty-portkey-litellm-openrouter - their own marketing comparison; Rust, 8ms P50 claim, OpenRouter’s 5.5% credit fee (published 2026-06-23)
https://www.truefoundry.com/blog/requesty-ai-pricing - competitor-authored critical guide; BYOK 0%, 30-day prompt retention, SOC 2 in progress (published 2026-09-11)
https://hn.algolia.com/api/v1/search?query=requesty.ai&tags=story - no meaningful Hacker News footprint found (queried 2026-09-26)

ai-gateway on tomrochette.com

OpenCode Zen

What it is
#

Status
#

Strengths
#

Cautions
#

Pricing
#

Price history
#

Compared to
#

Bottom line
#

Changes
#

See also
#

References
#

Requesty

What it is
#

Status
#

Strengths
#

Cautions
#

Pricing
#

Price history
#

Compared to
#

Bottom line
#

Changes
#

See also
#

References
#

ai-gateway on tomrochette.com

OpenCode Zen

What it is #

Status #

Strengths #

Cautions #

Pricing #

Price history #

Compared to #

Bottom line #

Changes #

See also #

References #

Requesty

What it is #

Status #

Strengths #

Cautions #

Pricing #

Price history #

Compared to #

Bottom line #

Changes #

See also #

References #

What it is
#

Status
#

Strengths
#

Cautions
#

Pricing
#

Price history
#

Compared to
#

Bottom line
#

Changes
#

See also
#

References
#

What it is
#

Status
#

Strengths
#

Cautions
#

Pricing
#

Price history
#

Compared to
#

Bottom line
#

Changes
#

See also
#

References
#