MiniMax H3 is now available on NanoGPT. Generate 2K video at 24 FPS with native stereo audio from text or an image. Create first-to-last-frame transitions, or guide subjects, motion, and sound with image, video, and audio references. Create with H3: https://nano-gpt.com/media?mode=video&model=minimax-h3

July's payment statistics are in. For the first time in a while, no coin had a discount at any point during the month, giving us a clean look at payment preferences. XMR led at 36.42%, followed by BTC at 20.44%. USDT and USDC combined for 18.22%, while XNO accounted for 5.48%. See the full month-by-month breakdown: https://nano-gpt.com/blog/crypto-payment-statistics

PixVerse v6 now supports reference-to-video on NanoGPT. Use up to 10 reference images to keep characters, products, scenes, and styles consistent. Or use up to 2 reference videos to guide motion, camera movement, and timing. Optional synchronized audio is also supported. Try reference-to-video: https://nano-gpt.com/media?mode=video&model=pixverse-v6

Demucs Stem Separation is now available on NanoGPT. Upload a stereo mix and separate it into vocals, drums, bass, guitar, piano, and other stems. Download each track as WAV or MP3 for remixes, karaoke, practice tracks, or cleaner audio edits. Separate a track: https://nano-gpt.com/media?mode=audio&model=fal-ai%2Fdemucs

NanoGPT's Benchmarks page now has dedicated leaderboards for Roleplay, Creative Writing, Coding, and Math. Each combines relevant benchmark sources into a weighted score, making it easier to compare models for the kind of work you actually want to do. Explore the new leaderboards: https://nano-gpt.com/benchmarks

We're opening an early look at a faster, more responsive NanoGPT chat experience. Long conversations stay smooth, research activity and sources are easier to follow, and Intelligent Search lets the model decide when another web lookup would improve its answer, within the limit you choose. You can also opt into tools such as location and spreadsheet creation. Enable it from the latest update in NanoGPT. You can return to the classic experience at any time. https://nano-gpt.com/

Ling 3.0 Flash and Ling 3.0 Flash Thinking are now available on NanoGPT! This 124B-parameter Mixture-of-Experts model activates about 5.1B parameters per token and is built for token-efficient coding, tool use, and production-scale agent workflows. Use Ling 3.0 Flash for faster direct answers, or Ling 3.0 Flash Thinking for harder reasoning tasks. Try Ling 3.0 Flash: https://nano-gpt.com/conversation?model=inclusionai%2Fling-3.0-flash Or Ling 3.0 Flash Thinking: https://nano-gpt.com/conversation?model=inclusionai%2Fling-3.0-flash%3Athinking

Qwen Image 3 is now available on NanoGPT! Generate images from Chinese or English prompts, render text accurately, or make precise edits using up to three reference images, all through one model. Great for posters, product shots, typography, and reference-driven edits. Try Qwen Image 3: https://nano-gpt.com/media?mode=image&model=qwen-image-3

Got thousands of AI requests that don't need instant answers? NanoGPT's Batch API runs chat completion jobs asynchronously for 50% less than synchronous requests. It is a strong fit for evals, classification, summarization, synthetic data, document processing, and image analysis. Upload a JSONL file, create a batch, monitor or cancel it, then download the results when ready. Manage batch jobs: https://nano-gpt.com/batch-api Read the guide: https://nano-gpt.com/blog/when-to-use-ai-batch-api

Poolside's Laguna S 2.1 and Laguna S 2.1 Thinking are now available on NanoGPT! The open-weights coding model has 118B total parameters with 8B active per token. It is built for long-horizon software engineering, tool use, and extended reasoning, with up to a 1M-token context window. Use the regular version for faster direct answers or Thinking for harder reasoning-heavy work. Both are included in the NanoGPT subscription. Try Laguna S 2.1: https://nano-gpt.com/conversation?model=poolside%2Flaguna-s-2.1 Try Laguna S 2.1 Thinking: https://nano-gpt.com/conversation?model=poolside%2Flaguna-s-2.1%3Athinking