Console
Independent AI API gateway

Buy Claude + DeepSeek + Qwen API Access — All Models, One Platform

One gateway key across frontier models — Claude for reasoning, DeepSeek for code, Qwen for value and vision. OpenAI-compatible, Anthropic-compatible and Responses endpoints behind a single custom base URL.

Create API key View quickstart
New accounts get $2.50 free test usage (valid for 7 days)

ThunderPeak is an independently operated API gateway. It is not affiliated with or endorsed by Anthropic, OpenAI or any model provider. Product and model names describe compatibility only.

CLAUDE
API
v1
Claude Opus 5
DeepSeek V4
Qwen 3.8 & Vision
platform usage packs
$11to$10 $55to$50 $230to$200

One gateway for Claude Code, Codex, Cursor and production API calls.

base_url: https://thunder-peak.com/v1
key: sk-tp-****
Choose a usage pack
Every major model Anthropic Claude DeepSeek Alibaba Qwen Moonshot Kimi Zhipu GLM OpenAI GPT / Codex

At a glance

What is ThunderPeak?

ThunderPeak fronts multiple upstream providers behind one OpenAI/Anthropic-compatible endpoint. Use the same balance for coding agents, IDE tools, SDKs and production workloads — with per-key limits and token-level usage in one dashboard.

View plans Read the quickstart Contact support

API formats

Anthropic Messages at /v1/messages, OpenAI-compatible chat at /v1/chat/completions, and Responses for Codex-class tools.

Built for tools

Custom base URL works with Claude Code, Codex CLI, CodeWhale, Cursor, Cherry Studio, VS Code extensions and any OpenAI SDK.

One balance

Credit packs never expire. Every request is metered by input, output, cache and thinking tokens.

8+
Models
3
API formats
4+
Upstream providers
1
Key · All models

Setup in minutes

Connect once. Work anywhere.

Point any tool that supports a custom base URL at ThunderPeak. No per-model accounts, no overseas cards required for your users — just one key.

Claude CodeCodex CLICursor CodeWhaleCherry StudioOpenAI SDKVS Code
One-prompt setup

Paste the ready-made config block into your tool and go.

1

Create a key

Generate one sk-tp key in your dashboard.

2

Pick a tool

Claude Code, Codex CLI, CodeWhale, Cursor or any OpenAI SDK.

3

Set the base URL

https://thunder-peak.com/v1 — start calling models.

01

One key, every provider

Claude for reasoning, DeepSeek for code, Qwen for value — routed automatically per model name.

02

Multi-protocol

Anthropic Messages, OpenAI chat and Responses endpoints from the same balance.

03

Streaming ready

SSE streams for coding agents, IDEs and production calls — billed the same as non-streaming.

04

Spend control

Per-key caps, request history and token-level usage clarity in your dashboard.

Native feel

Model access that feels native in your tools.

Use the same gateway key for agents, IDE extensions and production APIs. Keep setup simple, billing visible and control per key.

Models

One key, every frontier

Claude · GPT · DeepSeek · Qwen · Kimi · GLM — flagship discounts and value picks under one balance.

Browse all models
OpenAI GPT
GPT-5.5 · 5.6 · Codex

GPT-5.5 flagship reasoning · Terra / Sol for 1M-context work · Codex for agentic coding pipelines.

GPT-5.5 · flagship
Terra / Sol · 1M ctx
Codex · coding
Online
Claude
Anthropic · frontier trio

Opus 5 "frontier intelligence at half the price" · Sonnet 5 "the most agentic Sonnet" · Haiku 4.5 fastest & cheapest.

Fable 5 · 1M ctx
Opus 5 · 1M ctx
Sonnet 5 · 1M ctx
Haiku 4.5 · 200K
Online
Kimi
Moonshot · long context

K3 flagship with strong agentic and long-document performance; K2.7 Code tuned for engineering work.

K3 · flagship
K2.7 Code
Online
GLM
Zhipu · coding strength

GLM-5.3 top-tier coding gains; 5.3 Flash cheap and quick; 5.2 the high-value 1M-context option.

GLM-5.3
GLM-5.3 Flash
GLM-5.2 · 1M ctx
Online
DeepSeek
Open weights

V4-Pro rivals top closed-source models; V4-Flash is the economical workhorse. Native 1M context.

V4 Pro / V4 Flash
Online
Qwen
Alibaba · value & vision

3.8-Max 2.4T open MoE delivering 10+ day projects; 3.8-Flash 6B active with 1M context; VL for vision.

3.8 Max / Flash
Online

View full catalogue — Claude · GPT · DeepSeek · Qwen · Kimi · GLM →

Credit packs · no expiry

Simple usage-based pricing

Buy credit once, spend it only when API requests run. Every request is visible in your dashboard with token-level detail.

Starter balance
$11to$10

For daily agent work, hobby projects and trying every model.
Growth balance
$55to$50

For active teams, heavier coding sessions and production calls.
Free test bonus

New accounts get $2.50 of test usage after signup — verify the gateway before buying credits.

$2.50Free

What gets metered

Streaming is billed the same as non-streaming requests.

Input tokensOutput tokensCache writeCache readThinking tokens Buy credits View detailed pricing

Common questions

Short answers to what usually matters before buying an API key.

Buy a one-time credit pack; usage is deducted from your balance at the published per-model token rates. Credits never expire and there is no subscription.
New accounts receive $2.50 of free test usage after signup (valid for 7 days) so you can check the gateway before buying credits.
Claude Opus 5 / Fable 5 / Sonnet 5 / Haiku 4.5, DeepSeek V4 Pro / Flash, Qwen 3.8 Max / Flash and Qwen VL for vision. The live catalogue for your key is always GET /v1/models.
Sonnet 5 is the daily-driver balance of speed and quality; Opus 5 for the hardest reasoning; Haiku 4.5 for high-volume tasks. DeepSeek V4 Flash is the value pick for bulk code work.
OpenAI-compatible tools: https://thunder-peak.com/v1. Anthropic-compatible tools: https://thunder-peak.com.
No. ThunderPeak is an independently operated gateway. Model names describe API compatibility only, and your product should say the same to its own users.
Sign up in the console, generate a key, and pick the API format your tool needs. Full one-prompt setup is shown after signup.
The gateway is reachable worldwide via our edge network; actual performance depends on your local connectivity. Support is available in multiple languages.
Open setup docs

Ready to start building?

Get your API key and send your first request in under a minute.

Get API key Read documentation