# Inference: rejected and dead

> Model APIs that answer a completion request, keyless or behind a free key.

108 records: 40 usable, 26 trial, 20 paid only, 10 rejected, 12 dead. This page lists the rejected and dead records, each with the evidence that rejected it; the usable, trial and paid-only ones are on https://fshy.ajam.dev/t/inference.md. Each line: provider (its page in Markdown), record id (its JSON is https://fshy.ajam.dev/api/v1/record/<id>), verdict, free tier, identity, card, phone, evidence, stale date, parsed price, summary. In full: https://fshy.ajam.dev/t/inference/llms-full.txt.

## Rejected (10)

- [Baidu Qianfan ModelBuilder (ERNIE-Speed, ERNIE-Lite, ERNIE-Tiny presets)](https://fshy.ajam.dev/p/baidu.md) [`inference/baidu-qianfan-modelbuilder`](https://fshy.ajam.dev/api/v1/record/inference/baidu-qianfan-modelbuilder): fails-identity | free tier | national-id | no card | no phone | documented | stale on 2026-11-28. An announcement made the pay-as-you-go preset services ERNIE-Speed-8K, ERNIE-Speed-128K, ERNIE-Lite-8K, ERNIE-Lite-128K and ERNIE-Tiny-8K free of charge on...
- [Hack Club AI](https://fshy.ajam.dev/p/hackclub.md) [`inference/hack-club-ai`](https://fshy.ajam.dev/api/v1/record/inference/hack-club-ai): fails-identity | free tier | institution | no card | no phone | documented | stale on 2026-11-28. Every account can spend $3 per day across all endpoints and make 7,500 requests per 30 minutes; no credit card
- [ModelScope API-Inference](https://fshy.ajam.dev/p/modelscope-2.md) [`inference/modelscope-api-inference`](https://fshy.ajam.dev/api/v1/record/inference/modelscope-api-inference): fails-identity | free tier | national-id | no card | phone | documented | stale on 2026-12-04. Free API access to selected open-source LLM, multimodal and text-to-image models, described as a non-commercial, non-profit product.
- [Moonshot Kimi API Platform](https://fshy.ajam.dev/p/kimi.md) [`inference/moonshot-kimi-api`](https://fshy.ajam.dev/api/v1/record/inference/moonshot-kimi-api): fails-identity | free tier | country-phone | no card | documented | stale on 2026-11-28. An RMB 15 voucher granted on registration, valid for 3 months from issue, only for new users who register with a mainland China mobile number.
- [NLP Cloud](https://fshy.ajam.dev/p/nlpcloud.md) [`inference/nlp-cloud`](https://fshy.ajam.dev/api/v1/record/inference/nlp-cloud): fails-identity | free tier | email | no card | no phone | documented | stale on 2026-12-05. A prepaid Free plan at $0 with 2 parallel requests and variable per-model request rates, and a pay-as-you-go plan that credits $15 automatically; the...
- [OpenCode Zen free models](https://fshy.ajam.dev/p/opencode.md) [`inference/opencode-zen-free`](https://fshy.ajam.dev/api/v1/record/inference/opencode-zen-free): fails-identity | free tier | email | no card | no phone | measured | stale on 2026-12-05. The Zen pricing table lists 11 model ids at Free for input and output; 10 of them are also Free for cached read (Big Pickle, Pulse Alpha Free, GLM-5.3-Flash...
- [Tencent Hunyuan text generation API (China site)](https://fshy.ajam.dev/p/tencent.md) [`inference/tencent-hunyuan-api-cn`](https://fshy.ajam.dev/api/v1/record/inference/tencent-hunyuan-api-cn): fails-identity | free tier | national-id | no card | no phone | documented | stale on 2026-12-04. A one-time free resource package on first activation: 共100万 tokens shared across Hunyuan-a13b, Hunyuan-role-latest, the translation models and the vision...
- [g4f.space public provider routes (GPT4Free hosted)](https://fshy.ajam.dev/p/g4f.md) [`inference/g4f-space-public-routes`](https://fshy.ajam.dev/api/v1/record/inference/g4f-space-public-routes): fails-quota | free tier | anonymous | no card | no phone | reported | stale on 2026-11-12. Five routes (/api/groq, /api/ollama, /api/pollinations, /api/nvidia, /api/gemini) are listed as none required for an API key, but a chat call answered 402...
- [Public Ollama instances](https://fshy.ajam.dev/p/oneplus1.md) [`inference/public-ollama-instances`](https://fshy.ajam.dev/api/v1/record/inference/public-ollama-instances): fails-quota | anonymous | no card | no phone | reported | stale on 2026-11-12. Lists of bare IP:port endpoints running Ollama with no authentication.
- [uncensored-llm.sbs](https://fshy.ajam.dev/p/uncensored-llm.md) [`inference/uncensored-llm`](https://fshy.ajam.dev/api/v1/record/inference/uncensored-llm): fails-quota | free tier | anonymous | no card | no phone | measured | stale on 2026-11-28. 3 model ids, no key; no quota is published

## Dead (12)

- [DeepL API (Free plan)](https://fshy.ajam.dev/p/deepl.md) [`inference/deepl-api-free`](https://fshy.ajam.dev/api/v1/record/inference/deepl-api-free): dead | no free tier | card? | card? | phone? | documented. 500,000 characters per month for free on the DeepL API Free plan; the plan can no longer be purchased and has been replaced by the one-time 1,000,000...
- [Gemini CLI free tier (Gemini Code Assist for individuals)](https://fshy.ajam.dev/p/github-google-gemini.md) [`inference/gemini-cli-free`](https://fshy.ajam.dev/api/v1/record/inference/gemini-cli-free): dead | no free tier | email | no card | phone? | documented. Ended: Gemini CLI stopped serving the free Gemini Code Assist for individuals route on June 18, 2026, and free individual access moved to Antigravity CLI
- [GitHub Models inference](https://fshy.ajam.dev/p/github-2.md) [`inference/github-models-inference`](https://fshy.ajam.dev/api/v1/record/inference/github-models-inference): dead | no free tier | anonymous | no card | no phone | measured. No completion is returned; the endpoint answers 200 with a 4-byte body 'OK'
- [GLHF (glhf.chat)](https://fshy.ajam.dev/p/glhf.md) [`inference/glhf`](https://fshy.ajam.dev/api/v1/record/inference/glhf): dead | no free tier | email? | card? | phone? | measured. None now: the glhf.chat host answers HTTP 526 and the product is renamed Synthetic, a $30 a month subscription; the 2025-01-01 billing launch granted $10 of...
- [iFlow API (platform.iflow.cn)](https://fshy.ajam.dev/p/iflow.md) [`inference/iflow-api`](https://fshy.ajam.dev/api/v1/record/inference/iflow-api): dead | no free tier | email? | card? | phone? | documented. none: the vendor notice dated March 20, 2026 says the iFlow API service and model library close on April 17, 2026 together with iFlow CLI
- [KeylessAI](https://fshy.ajam.dev/p/keylessai-thryx-workers-dev.md) [`inference/keylessai`](https://fshy.ajam.dev/api/v1/record/inference/keylessai): dead | no free tier | anonymous | no card | no phone | measured. Documented as an OpenAI-compatible endpoint with zero API keys and zero signup routing to keyless upstream providers; the workers.dev host no longer...
- [Lambda Inference API](https://fshy.ajam.dev/p/lambda.md) [`inference/lambda-inference`](https://fshy.ajam.dev/api/v1/record/inference/lambda-inference): dead | no free tier | email? | card? | phone? | measured. none: the Inference API is winding down and its hosts do not resolve
- [LangSearch Rerank API](https://fshy.ajam.dev/p/langsearch.md) [`inference/langsearch-rerank-api`](https://fshy.ajam.dev/api/v1/record/inference/langsearch-rerank-api): dead | no free tier | email | no card | no phone | documented. none: the vendor states Rerank is no longer a documented product offering; the Free Plan covers web search with a daily token allowance
- [mlvoca.com free LLM API](https://fshy.ajam.dev/p/mlvoca.md) [`inference/mlvoca`](https://fshy.ajam.dev/api/v1/record/inference/mlvoca): dead | no free tier | anonymous | no card | no phone | measured. Documented as POST /api/generate on tinyllama and deepseek-r1:1.5b with no API key and no rate limit; when read, GET /api/generate answered 502 Bad Gateway...
- [Qwen Code OAuth free tier](https://fshy.ajam.dev/p/qwenlm-github-io.md) [`inference/qwen-code-oauth`](https://fshy.ajam.dev/api/v1/record/inference/qwen-code-oauth): dead | no free tier | email | no card | no phone | documented. Discontinued on 2026-04-15: Qwen OAuth is no longer a selectable /auth entry and new requests are rejected
- [uncloseai (unturf)](https://fshy.ajam.dev/p/unturf.md) [`inference/uncloseai`](https://fshy.ajam.dev/api/v1/record/inference/uncloseai): dead | no free tier | anonymous | no card | no phone | measured. No keyless access: GET /v1/models answers 401 public_beta_closed, saying the endpoint needs an API key and key registration is closed
- [ZeroEntropy zerank reranker and zembed-1 embeddings API](https://fshy.ajam.dev/p/zeroentropy.md) [`inference/zeroentropy-rerank-embeddings`](https://fshy.ajam.dev/api/v1/record/inference/zeroentropy-rerank-embeddings): dead | no free tier | email | no card | no phone | measured. none: the vendor sunset its hosted products after September 4th, 2026 and disabled new signups; the FAQ still advertises a free tier of unstated size.