Lumi API

Website Share on X

api.heang.top

A unified API gateway providing access to multiple LLM providers, billed by usage with no required subscription.

Models: 54 models
From: $0.160/M
Speed: 93 tok/s
Updated: 7/13/2026

Latency: 0.00 s
Created At: 4/25/2026
Recharge Rate: ¥1.00 per $1 quota

Features

DrawingTaskData ExportCheck-in

Login Methods

GitHubPasskey

API Endpoints

Endpoint 1
https://api.heang.top/v1
Endpoint 2
https://api.heang.top/v1/responses
Documentation
https://docs.heang.top
View model list, client setup, billing, and FAQs.
Endpoint 4Historical / Unverified
https://api.heang.top

Claim this provider

Verify ownership to unlock provider management features:

Edit provider name, content, and links
Get featured with priority traffic and visibility boost
Display a verified badge to build user trust

Leaderboard Rankings

Health

1.0%#324/716

Data as of Jul 13, 2026, 12:21 PMRankings are based on community-submitted tests and periodic health probes. Advisory only, not official data.

About Lumi API

Lumi API is a unified gateway that standardizes access to over 30 large language model (LLM) providers and services.

Key Features:

Unified API Endpoints: Provides endpoints compatible with popular standards for tasks like chat completions (/v1/chat/completions), embeddings (/v1/embeddings), image generation (/v1/images/generations), and audio transcription (/v1/audio/transcriptions).
Multi-Provider Access: Aggregates models from providers including OpenAI, Anthropic (Claude), Google (Gemini), Cohere, Moonshot AI, Zhipu AI, DeepSeek, and Grok, among others.
Usage-Based Billing: The service uses a pay-as-you-go model with no mandatory subscription.
Technical Integration: Users integrate by replacing the base URL in their API calls with Lumi API's gateway while maintaining their original endpoint paths.

Typical Use Cases:

Developers seeking a single API interface to access and switch between various LLMs.
Applications requiring multimodal AI capabilities (text, image, audio) through a consolidated service.

Health Check

100%Recent availability

History (72 pts)

PastNow

API Benchmarks & Pricing

Compare 49 model rows across audit recency, latest speed tests, throughput, latency, and per-token pricing.

View all 45 models

Model	Input ($/M)	Output ($/M)
doubao-embedding-vision-241215default	$1.80/M	$1.80/M
doubao-embedding-vision-250328default	$1.80/M	$1.80/M
doubao-embedding-vision-250615default	$1.80/M	$1.80/M
doubao-embedding-vision-251215default	$1.80/M	$1.80/M
doubao-seed-2.0-codedefault	$2.60/M	$2.60/M
doubao-seed-2.0-litedefault	$2.00/M	$2.00/M
doubao-seed-2.0-prodefault	$2.60/M	$2.60/M
doubao-seed-codedefault	$2.60/M	$2.60/M
glm-5.2-codingdefault	$2.70/M	$9.45/M
smart-coderdefault	$2.70/M	$10.80/M
step-image-edit-2default	$0.030/request	-
step-router-v1default	$1.60/M	$9.60/M
stepaudio-2.5-asrdefault	$3.00/M	$3.00/M
stepaudio-2.5-chatdefault	$12.00/M	$30.00/M
stepaudio-2.5-realtimedefault	$12.00/M	$84.00/M
stepaudio-2.5-ttsdefault	$0.700/request	-
gpt-5.6-terradefault	$2.70/M	$2.70/M
gpt-5.6-soldefault	$2.70/M	$2.70/M
glm-5.2default	$2.70/M	$9.45/M
qwen3.7-plusdefault	$2.70/M	$10.80/M
agnes-video-v2.0default	$1.50/request	-
agnes-video-v2.0vip	$1.50/request	-
agnes-image-2.1-flashdefault	$0.200/request	-
agnes-image-2.1-flashvip	$0.200/request	-
agnes-2.0-flashvip	$0.160/M	$0.160/M

Showing 25 of 49 model rows

Recent Test Records

Time	Model	Speed	Latency
Apr 25, 11:10 AM	gpt-5.4-mini	92.94 tok/s	4.06s

Similar API Provider Alternatives to Compare

Compare Lumi API alternatives against 6 nearby API providers using 323 LMSpeed signals across shared model coverage, pricing, benchmark speed, uptime, and free-model availability.

Provider	Why compare	Models	Free	Avg price	Speed	30d uptime
Lumi API api-heang-top A unified API gateway providing access to multiple LLM providers, billed by usage with no required subscription.	Current provider baseline	54	0	$0.160/M	93 tok/s	9610%
ChooseC API ipv4-beta-lm-studio ChooseC API is a unified AI model aggregation gateway supporting 260+ mainstream models including Claude, GPT, Qwen, DeepSeek, Kimi, and GLM with OpenAI, Claude, and Gemini compatibility.	Lower average pricing Faster measured speed Higher 30-day availability More free-model options Broader model coverage	335	9	$0.095/M	138 tok/s	9950%
Elysiver API elysiver-api Elysiver API appears to provide an OpenAI-compatible API gateway at elysiver.h-e.top. The public root is protected by a Cloudflare challenge during review.	Lower average pricing Faster measured speed Higher 30-day availability Broader model coverage	78	0	$0.014/M	119 tok/s	9940%
F2API api-f2api-com F2API is an enterprise AI gateway providing a unified OpenAI-compatible interface for accessing multiple large language models.	Faster measured speed Higher 30-day availability Broader model coverage	1,262	0	N/A	272 tok/s	9990%
Koyeb AI Gateway new-api-koyeb-app An OpenAI-compatible API gateway deployed on Koyeb, providing access to multiple AI models.	Higher 30-day availability More free-model options Broader model coverage	578	420	N/A	33 tok/s	9960%
WONG公益站 wzw-pp-ua WONG公益站 is a free, community-driven OpenAI-compatible API relay. It aggregates 70+ AI models from providers like OpenAI, Anthropic, DeepSeek, xAI, Moonshot, and Zhipu AI, offering free API quotas for developers and researchers.	Lower average pricing Higher 30-day availability Broader model coverage	140	0	$0.013/M	46 tok/s	9970%
ChooseC API ipv4-beta-kxcym-top-3001 ChooseC API is an OpenAI-compatible relay service hosted at ipv4-beta.kxcym.top, providing access to multiple AI models through standardized endpoints.	Lower average pricing More free-model options Broader model coverage	122	9	$0.095/M	51 tok/s	0%

Announcements

warning8/6/2026

Notice on Temporary Service Adjustments

Recently, some upstream channels have experienced fluctuations in quota, stability, and response speed. We are currently consolidating available models and backup channels. To avoid repeated failures, some unstable models will be temporarily taken offline or hidden.

Currently available models may be fewer than usual. We apologize for the inconvenience. We will prioritize restoring stable, reasonably priced models suitable for daily use. Your existing account, balance, and API Key remain unaffected.

If you encounter call errors, please keep the request time, model name, and request ID and contact the administrator. We will investigate and resolve as soon as possible.

success7/15/2026

Model Plaza Information Fully Updated

All 32 available models now have Chinese usage descriptions
Covers series including GPT, Qwen, GLM, Kimi, Doubao, MiniMax, Claude, and Agnes
Model prices, billing methods, channel priority, and API addresses remain unchanged
Existing API Keys do not need to be replaced

If the page still shows old content, please refresh and try again. For call errors, keep the request ID and contact the administrator.

default7/12/2026

Please keep your API Key secure and avoid sharing it publicly. Model availability and response speed are subject to upstream services. If you encounter errors, keep the request ID for troubleshooting.

FAQ

Why does the model occasionally respond slowly?

The model generation speed mainly depends on upstream load, context length, and output length. For streaming requests, it is recommended to set a longer client timeout.

How is the balance deducted?

The system charges based on the input and output tokens of the model and the corresponding multiplier. The actual deduction is based on the usage logs.

What should I do if the recharge does not arrive?

Do not make duplicate payments. Keep the payment order number and contact the administrator to verify the payment callback status.

Notes

Health checks: Scope: the 72-hour chart and recent availability measure API connectivity only. Each bar summarizes one hour of checks. Targets: LMSpeed tries the configured health check URL and provider status URL first, then API endpoints derived from known API hosts and recent speed-test base URLs. A website host is considered only when it looks like an API endpoint. Probe steps: each candidate goes through DNS lookup, TCP connection, TLS handshake for HTTPS, and an HTTP HEAD request with redirects followed. Probing stops after the first reachable candidate. Reachable criteria: every required network step must succeed. An HTTP response below 500 is treated as reachable, including 401 because it confirms that an authenticated API endpoint responded, except for statuses classified as blocked. Blocked results: HTTP 403, 429, 521, 525, and 530, plus detected WAF or Cloudflare challenges, are shown as blocked and excluded from availability calculations because LMSpeed cannot determine whether the API itself is down. Model availability: when a dedicated test key is configured, LMSpeed sends an authenticated GET request to a derived /models endpoint and compares returned model IDs with this provider's listed models. These per-model results appear in Models & Pricing and are not included in the provider connectivity percentage. Timeouts: TCP connection, TLS handshake, HTTP connectivity, and model requests each use a 20-second timeout. A full run can take longer when several candidates are tried. Frequency: a background worker checks all providers every 5 minutes by default. The 72-hour chart combines those samples into hourly bars, and the schedule may be changed by the service operator. Limit: automated samples are not an SLA and do not guarantee account quota, every model, every region, or successful completion requests. Check the provider's own status page before making operational decisions.
Domain Rating data is sourced from Ahrefs. It is a 0–100 backlink-based domain strength signal and does not measure API speed or reliability.
Announcements and FAQ are read from this provider's NewAPI status snapshot when available. LMSpeed stores the original content and optional English translations from the provider status source, then shows the localized fields on this page.