Model Spend Arena2ND ED.
provider · 2026-10-03

Synthetic

Official site · origin not confirmed

Plan

Synthetic (Standard)

Tier$/monthIncluded usage
Standard$30.00—

console only quota converts to units

As advertised. 500 peticiones por 5h y pack. Se venden por packs acumulables: 1 pack = $1/dia o $30/mes; N packs multiplican el limite. Se anuncian '3x mas limite que Claude'.

What that actually means. Peticiones por ventana de 5h, dichas en la tabla de precios. Estaba anotado como opaco: el dato SI estaba, pero solo aparece con el JavaScript ejecutado, y la comprobacion anterior se hizo con un lector de texto.

Where to check what is left. synthetic.new account dashboard

Models covered. kimi-k3, glm-5.3 (beta, 512k), glm-5.3-flash, deepseek-v4.1-flash (beta, 512k), qwen3.8-27b, glm-4.7-flash, nemotron-3-super-120b-a12b-nvfp4, gpt-oss-120b, nomic-embed-text-v1.5 (embeddings)

Runs open-source models in private datacenters with an OpenAI-compatible API (works with Roo, Cline, Octofriend, opencode). Privacy-first: it states it never trains on your data and never stores API prompts or completions. Carries Kimi K3 (AA coding 72) plus GLM-5.3-Flash, DeepSeek V4.1 Flash (beta) and Qwen3.8-27B — a lot of frontier open coding for a flat $30/month. The Standard plan allows 500 requests per 5h with 1 concurrent request per model.

checked 2026-10-02 · medium (landing page — $30/mo standard and Kimi K3 confirmed; exact quota and any referral reward not published) · source 1 source 2