Field notes on frontier open models.

Field notes on frontier open models — prices, benchmarks, and what we run in production.

2026-09-09 Why we built Wallaby — and where we're going Wallaby exists because buying access to frontier open-weight models is still harder than it should be. What we run, what we charge, and how we make reliability verifiable. PinnedCompany
2026-09-17 Kimi K3 in OpenCode: the zero-config setup (we're in the built-in model catalog) OpenCode pulls its provider catalog from models.dev — and Wallaby is in it. No base URL, no config file, no provider JSON. One environment variable and a model refresh, and Kimi K3 is selectable in the TUI. Set up and verified end to end on a real account. Tutorial
2026-09-17 One endpoint, one bill: running Kimi K3 for your whole dev team Give every developer on your team their own API key with a hard budget cap, all drawing from one prepaid balance — no seat licenses, no per-user subscriptions, no end-of-month surprise. Set up and verified on a real account. Tutorial
2026-09-17 Union Alpha Free in OpenCode: how to try the free stealth model (and what to know first) Union Alpha Free is a zero-cost, anonymously-provided model in OpenCode's built-in catalog — 262k context, no key required. Here is how to enable it, what we verified hands-on, and the honest caveats before you point it at real work. Tutorial
2026-09-14 The cheapest API is the one you don't have to redo Per-token price is the visible cost of an LLM. Rework is the bigger one. A short essay on why the unit that matters is cost per accepted output — and how to measure yours in an afternoon. Playbook
2026-09-13 Kimi K3 API pricing compared: 17 providers in one table Every major API provider serving Moonshot AI's kimi-k3 reasoning model, with input, cached-input and output prices per million tokens, plus a blended workload price that shows what coding agents actually cost on each. Guide
2026-09-13 AGENTS.md as a token budget: positive rules are not enough — you need a forbidden list AGENTS.md is now a Linux-Foundation-governed standard read by 30+ coding agents. Used well, it is a cost control, not just onboarding docs. Our production playbook — including the forbidden list at the center of it. Playbook
2026-09-13 Kimi K3 in Codex CLI: profile-file setup with an OpenAI-compatible endpoint Configure OpenAI Codex CLI to run on the kimi-k3 reasoning model through Wallaby Token's API, using Codex's profile file. Verified end to end with a real coding task. Tutorial
2026-09-12 Kimi K3 in VS Code Copilot via BYOK: custom endpoint setup, verified Point VS Code's Bring Your Own Key Custom Endpoint at an OpenAI-compatible API and run the kimi-k3 reasoning model inside Copilot Chat and agent mode. Full setup with working configuration. Tutorial
2026-09-12 Kimi K3 in Cline: 5-minute setup with an OpenAI-compatible endpoint Configure Cline to use the kimi-k3 reasoning model through Wallaby Token's OpenAI-compatible API. Working setup in under five minutes, verified end to end. Tutorial
2026-09-10 Kimi K3 in Hermes Agent: four lines of YAML, verified end-to-end Point Hermes Agent at the kimi-k3 reasoning model through Wallaby Token's OpenAI-compatible API. Verified end-to-end — chat and tool use both work. Tutorial
2026-09-09 Topping up your Wallaby Token balance: a step-by-step guide How to add credit to your Wallaby Token account with a card, Apple Pay, or Google Pay — plus the pitfalls we hit ourselves when testing. Guide
2026-09-09 Kimi K3 in Claude Code: 5-minute setup Configure Claude Code to run Moonshot AI's kimi-k3 model through Wallaby Token's Anthropic-compatible endpoint, verified end-to-end. Tutorial
2026-09-09 Kimi K3 in Cursor with a third-party endpoint: we tested it, so you don't have to We tried to run Kimi K3 in Cursor through an OpenAI-compatible API endpoint. Two hard blocks make it impossible today — here is exactly what we hit, and what works instead. Tutorial