Free tools for LLM developers.

LLM cost calculator

Estimate your monthly Claude, GPT, Gemini, or DeepSeek bill — with prompt caching, workload presets, and a shareable URL for every calculation.

Benchmark comparison

Side-by-side benchmark scores for Claude Opus 4.8, GPT-5.5, Gemini 3.5 Flash, DeepSeek V4 Pro, and more across coding, agentic, reasoning, math, and long-context dimensions.

Cost-efficiency leaderboard

Weekly auto-updated ranking — frontier LLMs by intelligence-per-dollar. Best $/SWE-Bench, $/MCP-Atlas, $/HLE across Claude, GPT, Gemini, DeepSeek.

API reference + OpenAPI + Postman

Full Anvat gateway API reference with downloadable OpenAPI 3.1 spec and Postman v2.1 collection. Anthropic and OpenAI compatible endpoints, one base URL.

Try Opus 4.8 on your code

Bring-your-own-key in-browser demo. Paste any code, run it through Claude Opus 4.8 (the model that found the Zcash bug) for security audit, refactor, tests, or explanation.

Video walkthroughs

Step-by-step videos of real AI engineering workflows — security audits, agent setups, cost-optimization patterns. Every video ships with a full transcript.

Model catalog

Per-model pricing tables, benchmark summaries, strengths/weaknesses, and a working code sample for every frontier model in the Anvat catalog.

Gateway comparison

Honest side-by-side comparisons of Anvat against OpenRouter, LiteLLM, Portkey, Helicone, Cloudflare AI Gateway, Vercel AI Gateway, and direct Anthropic.

Use case guides

Setup guides for Claude Code, Cursor, coding agents, and RAG pipelines — exact env vars, code snippets, and integration patterns.

Integrations catalog

Every IDE, agent runner, SDK, and framework that works with Anvat — Claude Code, Cursor, Zed, Continue, Cline, Aider, the official OpenAI and Anthropic SDKs, LangChain, LlamaIndex, LiteLLM, DSPy, Vercel AI SDK, Promptfoo. Copy-paste setup per tool.

Live gateway status

Browser-side health probe to api.anvat.app/healthz with round-trip latency, process uptime, published SLOs, and the incident log. No pre-rendered all-systems badge — a real fetch from your edge.

Public roadmap

Now / Next / Later, in priority order. Each item ships with a one-line rationale of who it's for and why it matters.

MCP + Claude Skills catalog

Curated MCP servers and Claude Skills verified end-to-end against Anvat-routed models. Filesystem, GitHub, Postgres, Slack, Stripe, Supabase, more. Copy-paste install per entry.

Code snippet library

Filterable copy-paste examples across Python, TypeScript, Go, Rust, cURL — Anthropic SDK, OpenAI SDK, LangChain, Vercel AI SDK, LiteLLM. Every snippet runs as-is.

FAQ — direct answers

Frequently asked questions covering pricing, supported models, SDK compatibility, MCP, privacy, performance SLOs. FAQPage JSON-LD for rich results.

Workload cost examples

10 anonymised real-shape AI workloads (Claude Code, RAG, support triage, batch classification) with direct list price vs Anvat metered vs prepaid effective. Math fully visible.

BYOK API playground

Paste your Anvat API key, pick a model (Claude / GPT-5 / Gemini / DeepSeek), send a prompt, stream the response in the browser. Key never touches our marketing host.

Site search

Static client-side search across every public surface — blog, models, snippets, FAQ, MCP, integrations, workloads, roadmap. Deep-linkable via ?q=.

中文 / Chinese hub

Localised landing for zh-CN — Anvat positioning, quickstart, FAQ, key resources translated. Self-contained for SEO + sharing in Chinese channels.