Releases·Tool Release
LinkTokenView launches Chinese LLM API aggregator in beta with OpenAI-compatible access to DeepSeek, Qwen, Kimi, GLM
LinkTokenView (联视未来) opened beta of an API aggregation platform offering one OpenAI-compatible endpoint for Chinese models including DeepSeek, Qwen, Kimi, Doubao, GLM, Baichuan, Spark, MiMo, and StepFun. A single key and shared token balance work across models with pay-as-you-go billing and automatic multi-channel failover. It targets developers swapping only a Base URL and supports agent tools like Codex and Claude.
23 engagement1 sourceTue, Sep 8, 2026, 12:30 PM
AI Overview
- LinkTokenView (联视未来) opened beta of an API aggregator offering one OpenAI-compatible endpoint across DeepSeek, Qwen, Kimi, Doubao, GLM, Baichuan, Spark, MiMo, and StepFun models
- One API Key works across all models with unified token balance and pay-as-you-go billing; Beta subscription plans sell at 1.9x discount pricing
- New users get a trial subscription worth 10 yuan (3-day validity); referrals earn 10% rebate on friends' top-ups
- Platform launched deepseek-v4.1-flash and supports agent tools including Codex, Claude, Harness, and WorkBuddy
- Multi-channel routing automatically fails over to backup channels on rate limits or errors and removes failed channels until recovery
Related
Social·Community
Community crowdsources examples of open-weight model failures vs frontier models
A Hacker News user is collecting concrete task examples where open-weight models (GLM, DeepSeek, Kimi, Qwen) failed while frontier models (Opus, Fable, GPT) succeeded, or vice versa, to test the claim that open models 6 months behind the frontier are good enough for most work. The thread provides a structured template for reporting failures and successes.
A Hacker News user is collecting concrete task examples where open-weight models (GLM, DeepSeek, Kimi, Qwen) failed while frontier models (Opus, Fable, GPT) succeeded, or vice versa, to test the claim that open models 6 months behind the frontier are good enough for most work. The thread provides a structured template for reporting failures and successes.
57 engagement1 sourcehackernewsFri, Jul 10, 2026, 06:02 PM
Products·launched
Mac pricing and local AI model fit checker across 15 countries
A free web tool that scrapes every buildable Mac configuration and shows the cheapest price across 15 countries, with filters for specs and budget. It also estimates which open-weight LLMs (Qwen, DeepSeek, GLM, Kimi) can run on each Mac's memory and rough token/s ceilings, plus compares against RTX 5090, DGX Spark, and Ryzen AI Max+. Built to save time for buyers comparing international prices and local AI capability.
A free web tool that scrapes every buildable Mac configuration and shows the cheapest price across 15 countries, with filters for specs and budget. It also estimates which open-weight LLMs (Qwen, DeepSeek, GLM, Kimi) can run on each Mac's memory and rough token/s ceilings, plus compares against RTX 5090, DGX Spark, and Ryzen AI Max+. Built to save time for buyers comparing international prices and local AI capability.
0 engagement1 sourceredditWed, Aug 26, 2026, 08:28 PM
Products·launched
Linknux — unified OpenAI-compatible API for multiple LLM providers
Linknux is a unified AI API that lets developers call models from OpenAI, Anthropic, Google, DeepSeek, Qwen and others through a single API key, endpoint, and billing account. It is OpenAI-compatible, so existing OpenAI SDKs work by changing the base_url and model name. It targets developers who want to avoid managing separate accounts, keys, SDKs, and billing per provider.
Linknux is a unified AI API that lets developers call models from OpenAI, Anthropic, Google, DeepSeek, Qwen and others through a single API key, endpoint, and billing account. It is OpenAI-compatible, so existing OpenAI SDKs work by changing the base_url and model name. It targets developers who want to avoid managing separate accounts, keys, SDKs, and billing per provider.
1 engagement1 sourceredditTue, Sep 1, 2026, 02:05 AM
Products·launched
Liquid Inference – LLM inference marketplace with auto-routing by price
Liquid Inference is a marketplace where LLM model providers compete on price, paired with an auto-router that picks the lowest-cost option for a preset quality level. It targets developers who want to cut inference costs without manually comparing providers. Built by a team with a trading-systems background.
Liquid Inference is a marketplace where LLM model providers compete on price, paired with an auto-router that picks the lowest-cost option for a preset quality level. It targets developers who want to cut inference costs without manually comparing providers. Built by a team with a trading-systems background.
2 engagement1 sourcehackernewsFri, Oct 9, 2026, 02:52 PM
Products·launched
MCP server for Claude Code to delegate tasks to other LLMs
An MCP server that allows Claude Code to delegate tasks to other models like GPT-5.6, DeepSeek, GLM, and local Qwen. The creator benchmarked all models against Claude itself with 198 runs and hidden tests. It solves the problem of using multiple models from a single interface for comparison or specialized tasks.
An MCP server that allows Claude Code to delegate tasks to other models like GPT-5.6, DeepSeek, GLM, and local Qwen. The creator benchmarked all models against Claude itself with 198 runs and hidden tests. It solves the problem of using multiple models from a single interface for comparison or specialized tasks.
298 engagement1 sourceredditMon, Jul 20, 2026, 06:25 PM