Provides API services with a model marketplace and developer tools.
https://metahk.zenymes.comLow-latency, high-concurrency batch image analysis
Verify ownership to unlock provider management features:
Meta API offers API services including a model marketplace for accessing various models, a console for management, and documentation. It supports developer tools and integration capabilities. Use cases include building applications that require model access and API management.
Compare 9 model rows across audit recency, latest speed tests, throughput, latency, and per-token pricing.
| Model | Input ($/M) | Output ($/M) | Audit | Speed | Latency |
|---|---|---|---|---|---|
deepseek-ai/deepseek-v3.2default | $73.97/M | $73.97/M | — | — | — |
| $0.367/M | $1.47/M | — | — | — | |
deepseek-v3.2-expdefault | $73.97/M | $73.97/M | — | — | — |
kimi-k2-thinkingdefault | $73.97/M | $73.97/M | — | — | — |
moonshotai/kimi-k2-thinkingdefault | $73.97/M | $73.97/M | — | — | — |
| $0.367/M | $1.47/M | — | — | — | |
MiniMaxAI/MiniMax-M2default | $0.367/M | $1.47/M | — | — | — |
claude-haiku-4-5default | $73.97/M | $369.86/M | — | — | — |
gemini-2.5-flash-imagedefault | $73.97/M | $616.44/M | — | — | — |
Showing 9 of 9 model rows
Compare Meta API alternatives against 6 nearby API providers using 236 LMSpeed signals across shared model coverage, pricing, benchmark speed, uptime, and free-model availability.
| Provider | Why compare | Models | Free | Avg price | Speed | 30d uptime |
|---|---|---|---|---|---|---|
| Meta API meta-api Provides API services with a model marketplace and developer tools. | Current provider baseline | 34 | 0 | $31.81/M | N/A | 99.8% |
| N1N api-n1n-ai N1N provides API access to a wide range of AI models including GPT-4, Claude 3, Gemini, and others for text, image, and video generation. |
| 177 | 10 | $6.60/M | 90 tok/s | 99.8% |
| 180txt API 180txt-cn 180txt API provides an OpenAI-compatible API relay for multiple AI models. |
| 37 | 8 | $31.07/M | 42 tok/s | 14.8% |
| IXIOCCAPI ixioccapi IXIOCCAPI is a unified API gateway for large language models, providing standardized endpoints for accessing multiple AI model providers. |
| 408 | 9 | $55.54/M | N/A | 99.6% |
| CaMeL AI api-kr777-top CaMeL AI provides an OpenAI-compatible API gateway with extensive model coverage and pricing options. |
| 194 | 6 | $79.70/M | 93 tok/s | 99.2% |
| WAADRI new-waadri-top WAADRI runs a unified AI model gateway that exposes aggregated model access through OpenAI-, Claude-, and Gemini-compatible interfaces. |
| 192 | 710 | N/A | N/A | 5.2% |
| LLM API llm-api Provides API access to large language models for developers. |
| 82 | 71 | N/A | N/A | 63.4% |
2025/7/10 Added the latest model grok-4!!!
The latest Grok 4 is xAI's latest reasoning model, featuring a 256k context window, supporting parallel tool calls, structured output, and image and text input.
2025/06/18 Launched the official version of gemini-2.5-pro model
Multimodal, excellent for image analysis, file analysis, and code programming (official version)
2025/06/06 New addition: the latest Gemini model gemini-2.5-pro-preview-06-05 Google's latest advanced AI model, designed for complex reasoning, programming, mathematics, and scientific tasks. It has 'thinking' capabilities, enabling it to reason more accurately and handle nuanced contexts in responses. Gemini 2.5 Pro excels in multiple benchmarks, ranking first on the LMArena leaderboard, reflecting its excellent human preference alignment and complex problem-solving abilities.
Multimodal, excellent for image analysis, file analysis, code programming, etc. 0.03 per call
1. Check if the token's authorized quota is sufficient; it is recommended to set it to 'unlimited quota'. 2. There might be a few minutes of system cache after recharging; try to use the same API endpoint for both API calls and recharging.
As long as the application supports binding an API key and API address (baseUrl), it can generally be used. Our API interface is compatible with most third-party applications. This site also provides binding tutorials for some applications; you can visit the Help Center page to view them.
Usage is charged based on chat word count (questions + replies). Since each person uses different models, question frequency, and question length, it is recommended to check the Pricing page to calculate usage.
1. Some users may experience intermittent instability due to DNS pollution, corporate environments, network operators, or educational network environments. 2. Regional restrictions: some areas may block overseas unregistered websites, listing them as untrusted sites. 3. When OpenAI servers are under high load, they may throttle speeds, which is relatively rare. 4. Using an unstable VPN can cause API access instability. (High probability) It is recommended to use without a VPN. 5. Frequent network switching in hotspot or distributed WiFi environments. (High probability) You can try changing to an accessible address based on the API information published on our console page.
Because AI models are named with version numbers only after training or upgrades are completed. The model's knowledge base does not contain information about its true version. Please verify GPT4 using the methods provided on this site.
1. [429] Current group default upstream model xxx is saturated. A: This indicates that the model's rate is at full capacity; please try again later. 2. Current group default has no available channel for model xxx. A: This indicates that the model has been automatically disabled due to a fault; it usually self-recovers within a few minutes. If it remains unavailable, manual intervention by an administrator is required. 3. [401], [403] errors occur! A: [401] Token error, API address not bound, etc.; [403] Insufficient balance, insufficient permissions, token model restrictions, etc.
A reverse model is like a 'cracked' version of an AI model. It allows us to bypass official AI channels (such as the ChatGPT website) and use AI services directly through program interfaces. However, this 'cracked' version has a significant drawback - it is not stable. It is like 'playing hide and seek with the official'; once the official updates the system or strengthens restrictions, these cracked solutions may fail and require continuous improvement to remain usable.
After the balance is used up, you can recharge online on the Wallet Recharge page of this site, or enter a recharge redemption code to exchange for balance. If your token is not set to 'unlimited quota', after recharging, you also need to check if the token's balance is sufficient to avoid affecting normal use. After recharging, you do not need to replace the token; just ensure that the token's authorized quota is sufficient.
https://metamrb.zenymes.comHigh-bandwidth, high-concurrency batch image analysis
https://chatshare.zenymes.comHigh-bandwidth backup route
Rankings are based on community-submitted tests and periodic health probes. Advisory only, not official data.