分类筛选
找到 1605 个 Demo
FREE LLM API ACCESS: THE COMPLETE LIST 21.1k star...
FREE LLM API ACCESS: THE COMPLETE LIST 21.1k stars and counting this repo tracks every provider giving away LLM inference for free → OpenRouter: 50 req/day free, 1000/day with $10 lifetime topup → Google AI Studio: Gemini 3 Flash + Gemma models, no cost → Cerebras: gpt-oss-120b at up to 30 req/min → Groq: Llama, Whisper, gpt-oss models, fast inference → Cloudflare Workers AI: 10,000 neurons/day free → GitHub Models: tier-based access via Copilot subscription Bonus: trial credits from Fireworks, Baseten, Nebius, SambaNova, and more for testing before you commit Built and maintained by @cheahjs https://t.co/6tjwT80Qlo
To fellow Tech people, My Company has given Micro...
To fellow Tech people, My Company has given Microsoft github copilot access. Has the below models: ▪︎ GPT-5 Mini ▪︎ GPT 5.4 ▪︎ Gemini 2.5 Pro ▪︎ Gemini 3.5 Flash ▪︎ Claude Sonnet 4.6 ▪︎ Claude Opus 4.6 What models do you have ? Repost and Answer ✍️
Gemini 3.5 Pro Leaks - Gemini 3.5 Pro is current...
Gemini 3.5 Pro Leaks - Gemini 3.5 Pro is currently targeting a July 17 launch, with the extra time being used for a new pretraining run instead of reusing the older Gemini 2.5 Pro base. - Frontend generation sees a major jump better design taste, cleaner UI, stronger SVG generation, and much more polished one shot frontend results - It's also much better at one shot game generation, including sound effects and music, though Fable still has the better overall. - Google is also preparing a Nano Banana Pro model based on the new 3.5 Pro foundation, expected to compete more directly with GPT-Image 1.
ONE GITHUB REPO has every permanently free AI API...
ONE GITHUB REPO has every permanently free AI API with real rate limits - no trials that expire, no cards needed, just models that stay free forever 🤯 not signup-credit stuff that runs out in a week, these are the tiers that stay free what's inside: - free llm apis with their exact rate limits - which ones are openai-sdk compatible - speed tiers and the free model list for each - cloud provider free tiers in one place - trial credit providers listed separately so you know what's permanent it covers google ai studio, openrouter, groq, cerebras, nvidia nim, github models, mistral, huggingface, cloudflare workers, cohere and dozens more - 24k+ stars, updated constantly how to use it: step 1: go to https://t.co/gPyF2dycyp > scroll down to "free providers" - these are the permanent ones step 2: pick a provider from the list > google ai studio - gemini 2.5 flash, free forever, no card > groq - llama, qwen, nemotron models, 30 req/min free > cerebras - fastest inference, 30 req/min, no card > openrouter - hermes, nemotron, qwen, poolside models step 3: copy the base url from the readme > each provider has its endpoint listed > most are openai-compatible - just swap the url and key step 4: paste into cursor, claude code, aider, hermes, any tool > set the base url to the provider's endpoint > paste your free api key (if needed) > pick a model name from the list > start sending requests - no billing setup required step 5: check the "trial credits" section for bonus usage > fireworks, baseten, nebius, hyperbolic - $50-100 free trial credits > use these after you drain the permanent free tiers one repo replaces hours of scrolling through pricing pages
Gemini 2.5 Pro and Gemini 3 Flash will be deprecat...
Gemini 2.5 Pro and Gemini 3 Flash will be deprecated in all Copilot experiences on July 31, 2026. • Update workflows to Gemini 3.1 Pro or Gemini 3.5 Flash before the deprecation date https://t.co/5amUNDm03r
22日，百度闷声干了票大的！最新开源的Unlimited OCR 总参数3B，实际激活仅500M—...
22日，百度闷声干了票大的！最新开源的Unlimited OCR 总参数3B，实际激活仅500M——放在大模型时代几乎是个零头。 但就是这个小到离谱的模型，在OmniDocBench v1.5上拿下93.23%的综合分，v1.6更是达到93.92%，直接刷新了端到端SOTA。 什么概念？v1.5同台竞技的选手里，235B的Qwen3-VL拿了89.15，72B的Qwen2.5-VL拿了87.02，不公布参数量的Gemini-2.5 Pro也只有88.03。激活参数不到它们零头的选手，反手把它们全甩了。 更离谱的是，它还干了件之前没有OCR模型干成过的事：一口气解析40多页文档，不失忆、不降速，一次推理从第一页读到最后一页。 目前，模型和代码都已同步上线GitHub和HuggingFace。 https://t.co/yuU7E121yW https://t.co/8YMkjoosB5
🚨 Gemini Pro 1 Year Giveaway! 🚨 ✨ Gemini 2.5 Pr...
🚨 Gemini Pro 1 Year Giveaway! 🚨 ✨ Gemini 2.5 Pro 💰 1000 AI Credits/Month 🎬 Veo 3 + Flow AI + Whisk 📚 NotebookLM 📩 Gmail & Docs Integration To Enter: 1️⃣ Follow @expertwith_AI 2️⃣ Like & RT 3️⃣ Comment “ Send ” ⏳ 48 Hours Only. https://t.co/r5zUMXaqrw
THIS GUY BUILT AN AI PIPELINE THAT MAKES INFLUENCE...
THIS GUY BUILT AN AI PIPELINE THAT MAKES INFLUENCERS DANCE AND NOW NIGHTLIFE BRANDS PAY HIM $1,800 PER CHARACTER. He got tired of watching every AI video account post models that stand still and blink so he built a 7-node system that outputs full-body dance sequences where the face stays locked and the movement actually looks like someone filmed it on a phone at 2am. Here's the exact breakdown: → Claude writes a character DNA file with 32 parameters before the first frame gets rendered. Not just face structure. Full body proportions, shoulder-to-hip ratio, arm length, the way weight shifts from one foot to the other. If the body DNA is not locked the limbs start stretching mid-sequence and the whole thing falls apart by second three. → Flux 1.1 Pro generates the base still from that DNA plus a locked moodboard reference. Club lighting, warm skin tones, motion blur on the background but sharpness on the face. The negative prompt runs 42 exclusion terms because nightlife shots have more failure modes than daylight ones. → 3 Gemini 2.5 Pro nodes correct isolated face zones on the base render. Lip geometry, iris depth, nose pore density. Under club lighting the AI defaults to smoothing everything into plastic so the correction pass has to fight the model's instinct to flatten texture. → Kling 2.6 animates the locked still into a dance sequence but only after a motion reference track is loaded. Not a text prompt saying "dance." An actual choreography reference that maps beat timing to body position. Text-prompted dance looks like a puppet. Reference-mapped dance looks like a person. → ElevenLabs drops a voice layer that matches the character's energy so the final clip works as a story reply or a live shopping teaser. The voice is cloned from a 90-second sample and tuned to sound like it belongs to that face under that lighting in that room. → The finished clip exports vertical 9:16 with native club atmosphere intact. No post-production color grade needed because the lighting was baked into the prompt architecture from step one. 12 finished clips per character per batch. [VIDEO INSERT / AI dance clip / club setting / 0:08] The key move 95% of people skip: you cannot text-prompt a dance. You have to feed a motion reference. Everyone types "woman dancing in a nightclub" and gets a figure that sways side to side like a screen saver. The hips rotate but the feet stay glued. The system loads an actual choreography reference that maps every beat to a body position. That one step is the difference between a clip that gets scrolled past and a clip that gets saved. Nightlife venues and event brands now pay $1,800 per character build and $349 a month to keep that AI performer as a permanent content asset for stories, reels, and event promo. Clubs that switched fill guest lists 22% faster because the content never waits for a photographer to show up on a Saturday night.
AI Timeline 2020: GPT-3 2021: GitHub Copilot 202...
AI Timeline 2020: GPT-3 2021: GitHub Copilot 2022: ChatGPT, Midjourney, Stable Diffusion 2023: GPT-4, Claude, Llama 2024: Sora, Veo, Claude 3.5 2025: GPT-5, Gemini 2.5, Grok 4, Claude 4 2026: Claude Fable 5, Mythos 5 2027:???
I've noticed recently that the Gemini 2.5 Flash-Li...
I've noticed recently that the Gemini 2.5 Flash-Lite model is producing a lot of 503 errors. Around 30% of my daily requests are failing. The reason is simple: model overloaded. Gemini 2.5 Flash is more expensive, so I'm not sure I'm ready to switch yet, but I don't really see any other option for now. Curious how everyone else is dealing with this.
Google is attacking the biggest barriers in genera...
Google is attacking the biggest barriers in generative media: waiting time and cost. Google introduced Nano Banana 2 Lite and Gemini Omni Flash for the Gemini API and Google AI Studio. Nano Banana 2 Lite is built for fast, low-cost image generation. It produces a 1K image in under four seconds at roughly $0.034, while still scoring competitively against larger image models. It reached an Elo score of 1251 for image generation and 1308 for image editing, placing it close to Nano Banana 2 while running five times faster. Gemini Omni Flash focuses on high-quality video generation and conversational editing. It delivers state-of-the-art video-editing performance at $0.10 per second, matching the price of Veo 3.1 Fast.
ICYMI: our June AI recap ⬇️ 🍌 Unveiled Nano Ban...
ICYMI: our June AI recap ⬇️ 🍌 Unveiled Nano Banana 2 Lite, our fastest and most cost-efficient Gemini Image model yet 💻 Introduced Gemma 4 12B, our latest open model that brings smart AI agents directly to your laptop in a single streamlined system to give you privacy without sacrificing speed ⚡ Brought Gemini Omni Flash to APIs in public previews so developers and enterprises can build custom, dynamic video workflows 🔊 Released the new Google Home Speaker, which we built with Gemini to have more natural and fluid conversations and to handle multiple requests at once 📓 Upgraded @NotebookLM with advanced reasoning, a secure cloud computer for code execution and the ability to generate charts, spreadsheets and slide decks for Google AI Ultra subscribers and specific Workspace accounts 📱Rolled out Android 17 with new features designed for multitasking, enhanced security, including lost phone biometric locking, and new personalization options 🗣️ Launched Gemini 3.5 Live Translate, which enables near-real-time speech-to-speech translation across 70+ languages by preserving natural intonation and eliminating pauses 📚 Rolled out study notebooks to the @GeminiApp to help users reach their learning goals by analyzing uploaded class notes and quiz results to create personalized lessons and track progress 🌪️ Shared our updated weather models that use AI to help communities stay safe by predicting climate crises, like river floods, wildfires and cyclones