Mastodawn

Your image generator has never seen today. It was trained months ago, maybe longer, and everything it draws comes from that frozen snapshot of the world.

Gen-Searcher does something none of the mainstream tools do. Before it draws a single pixel, it goes and looks things up. It searches the web. It browses sources. It pulls visual references. Then it generates. The result is an image grounded in actual current information.

https://firethering.com/gen-searcher-open-source-image-generation/

#opensource #ai #llm #genai #agent #trending

Gen-Searcher: An Open Source AI That Searches the Web Before Generating Images - Firethering

Your image generator has never seen today. It was trained months ago, maybe longer, and everything it draws comes from that frozen snapshot of the world. Ask it to generate a current news moment, a product that launched last month, or anything that requires knowing what's happening right now and it fills in the gaps with a confident guess. Sometimes that guess is close. Often it isn't. Gen-Searcher does something none of the mainstream tools do. Before it draws a single pixel, it goes and looks things up. It searches the web. It browses sources. It pulls visual references. Then it generates. The result is an image grounded in actual current information. It's open source, the weights are on Hugging Face, and the team released everything including code, training data, benchmark, the lot.

Firethering

Firethering 1d ago

File Converter Pro works like a simple offline converter. You drop files in, pick what you want, and it converts everything locally. No uploads on any server.
https://firethering.com/file-converter-pro-offline-file-converter-images-audio-video-documents/

#opensource #fileconverter #utilities #software #trending

File Converter Pro offline file converter for images audio video and documents - Firethering

Most file converters still push you to upload your files somewhere. Even for basic stuff like changing a PDF or converting an image. It works, but it’s not something you feel great about, especially with random files. File Converter Pro works like a simple offline converter. You drop files in, pick what you want, and it converts everything locally. No uploads or any server. The UI isn’t just functional, it actually looks like someone cared. Smooth startup, proper dark mode, small touches that make it feel like a real app instead of a side project. There’s also some extra stuff like stats and achievements. Sounds gimmicky, but it kind of works. You start noticing how often you use it. It’s not lightweight though. And if you want audio or video conversions, you’ll need FFmpeg. But once that’s sorted, you’re done setting things up.

Firethering

Firethering 2d ago

M2.7 is now available on HuggingFace with weights you can download and deploy.
https://firethering.com/minimax-m2-7-agentic-model/
#minimax #opensource #ai #agent #llm #genai #technews

MiniMax M2.7: The Agentic Model That Helped Build Itself - Firethering

MiniMax handed an internal version of M2.7 a programming scaffold and let it run unsupervised. Over 100 rounds it analyzed its own failures, modified its own code, ran evaluations, and decided what to keep and what to revert. The result was a 30% performance improvement with nobody directing each step. That is not a benchmark result. That is a different way of thinking about how AI models get built. M2.7 is now available on HuggingFace with weights you can download and deploy. NVIDIA is offering free API access if you want to try it without the hardware overhead. The license has a commercial limitation worth knowing about, we will get to that.

Firethering

Firethering 3d ago

Most AI models are what they appear to be. A 12B parameter model uses 12B parameters. What you see is what runs.

Marco MoE does not work that way. Alibaba built two models, Marco Nano and Marco Mini, that carry billions of parameters but wake up only a tiny fraction of them for each request. Marco Nano activates 0.6B out of 8B. Marco Mini activates 0.86B out of 17.3B. Less than 5% of either model actually works.

https://firethering.com/marco-moe-nano-mini/
#opensource #ai #alibaba #moe #huggingface #llm #genai

Marco MoE Uses 5% of Its Parameters but Outperforms Models 3× Its Size - Firethering

Most AI models are what they appear to be. A 12B parameter model uses 12B parameters. What you see is what runs. Marco MoE does not work that way. Alibaba built two models, Marco Nano and Marco Mini, that carry billions of parameters but wake up only a tiny fraction of them for each request. Marco Nano activates 0.6 billion out of 8 billion. Marco Mini activates 0.86 billion out of 17.3 billion. Less than 5% of either model is actually working at any moment. The part that makes this worth paying attention to is what that 5% manages to do against models running at full capacity.

Firethering

Firethering 3d ago

Most AI voice tools give you two options. Clone an existing voice or pick from a list of defaults. If neither works for what you need, you are stuck.

VoxCPM2 adds a third option. You describe what you want. A young woman, gentle tone, slightly slow pace. A deep male voice with a formal cadence. Whatever you can put into words, it generates from scratch, no recording needed.
https://firethering.com/voxcpm2-voice-cloning/
#tts #ai #trending #genai #opensource

VoxCPM2 lets you create voices just by describing them and it is open source - Firethering

Most AI voice tools give you two options. Clone an existing voice or pick from a list of defaults. If neither works for what you need, you are stuck. VoxCPM2 adds a third option. You describe what you want. A young woman, gentle tone, slightly slow pace. A deep male voice with a formal cadence. Whatever you can put into words, it generates from scratch, no recording needed. That alone would make it interesting. But it also does voice cloning, supports 30 languages without needing a language tag, outputs 48kHz audio, runs on 8GB of VRAM, and ships under Apache 2.0. The whole thing is two billion parameters and installs with a single pip command. I tried the audio samples and the results are genuinely good. Not fully human, but natural enough that you stop noticing the model and start paying attention to what it is saying. Mixed languages, different emotions, and you can steer all of it.

Firethering

Firethering 5d ago

Muse Spark launched yesterday under Meta Superintelligence Labs, a new internal division Meta quietly formed by bringing together researchers from Google DeepMind and other frontier labs. It is natively multimodal, supports multi-agent reasoning, and is available right now at meta.ai.

https://firethering.com/meta-muse-spark-multimodal-ai/

#meta #ai #musespark #trending #llm

Meta’s Muse Spark: A Closed Bet on Multimodal, Multi-Agent AI - Firethering

Meta has a new AI model and for the first time in years it is not called Llama. Muse Spark launched yesterday under Meta Superintelligence Labs, a new internal division Meta quietly formed by bringing together researchers from Google DeepMind and other frontier labs. It is natively multimodal, supports multi-agent reasoning, and is available right now at meta.ai. It is also not being released as open weights. That last part is worth sitting with for a second. Meta built one of the most trusted brands in open source AI through Llama. Developers built on it, researchers published with it. Muse Spark continues none of that. No weights, no HuggingFace release, private API preview only. What you get instead is a genuinely capable multimodal model with some benchmark numbers that are hard to ignore and a new reasoning mode called Contemplating that puts it in conversation with Gemini Deep Think and GPT Pro. Whether that trade is worth it depends entirely on what you were using Meta AI for in the first place.

Firethering

Firethering 6d ago

ZhipuAI ran GLM-5.1 on a vector database optimization problem and let it go for 600 iterations. It did not run out of ideas. At iteration 50 it was sitting at roughly the same performance as the best single-session result any model had achieved. By iteration 600 it had reached 21,500 queries per second. The previous best was 3,547.

The model is MIT licensed & Available on HuggingFace .

Here's what this model can do & who is it for
https://firethering.com/glm-5-1-open-source-agentic-model/
#opensource #ai #glm51#trending #llm

GLM 5.1: The open source model that gets better the longer you run it - Firethering

Give an AI agent a hard problem and it usually figures out the easy wins fast. After that, more time does not help. It just sits there, trying the same things. ZhipuAI ran GLM-5.1 on a vector database optimization problem and let it go for 600 iterations. It did not run out of ideas. At iteration 50 it was sitting at roughly the same performance as the best single-session result any model had achieved. By iteration 600 it had reached 21,500 queries per second. The previous best was 3,547. That gap is not incremental improvement. It is a different category of result. GLM-5.1 is open source, MIT licensed, and the weights are on HuggingFace right now. It works with Claude Code, vLLM, and SGLang. If you are building anything that runs agents over long tasks, this one is worth understanding.

Firethering

Firethering Apr 7

PrismML released Bonsai 8B & the whole model, weights and all, fits in 1.15 GB. For context, the standard FP16 version of a comparable 8B model sits at around 16 GB. Bonsai beats or matches several of them on benchmarks
https://firethering.com/bonsai-8b-1bit-llm/

#ai #opensource #llm #trending #genai #huggingface

Bonsai 8B: A 1-Bit LLM That Delivers 8B-Class Performance at 1/14th the Size - Firethering

Nobody expected a 1.15 GB model to score competitively against full precision 8B models. That is not how this usually goes. PrismML released Bonsai 8B last month and the headline number is almost absurd. The whole model, weights and all, fits in 1.15 GB. For context, the standard FP16 version of a comparable 8B model sits at around 16 GB. Bonsai beats or matches several of them on benchmarks while being 14 times smaller. It runs on a phone. There is literally an iPhone build. I want to be clear that these numbers come from PrismML's own evaluations, not independent third party testing. But even with that caveat, this is worth paying attention to.

Firethering

Firethering Apr 7

Running a local LLM usually means a Python environment, CUDA drivers, and at least one Stack Overflow tab open before you’ve even started. llamafile skips all of that. Mozilla ai packaged the whole runtime like model weights and everything into a single executable. On Windows you rename it to .exe. On Mac or Linux you chmod +x it. That’s the setup.
https://firethering.com/llamafile-run-ai-models-locally-one-file/

#opensource #ai #llamacpp #llama #genai #huggingface

Llamafile: Run AI Models Locally on Your PC with Just One File - Firethering

Running a local LLM usually means a Python environment, CUDA drivers, and at least one Stack Overflow tab open before you've even started. llamafile skips all of that. Mozilla.ai packaged the whole runtime like model weights and everything into a single executable. On Windows you rename it to .exe. On Mac or Linux you chmod +x it. That's the setup.

Firethering

Firethering Apr 6

Now you can run AI models like Gemma 4 Directly in your phones easily with Google's Open Source App Google AI Edge Gallery , It supports multiple AI models you download in the app itself from Hugging face and it work fully Offline.

https://firethering.com/google-ai-edge-gallery-offline-llm-app/

#opensource #google #gemma #ai #trending #lllm

Google AI Edge Gallery: Run LLMs Offline on Your Phone

Google AI Edge Gallery lets you run open-source LLMs straight on your phone. No cloud. Once you download the models, you're offline. You get chat, image analysis, audio transcription, prompt testing. All on-device. Newer models like Gemma 4 mean better reasoning and multimodal stuff on mobile hardware. It’s more like a sandbox where you can test, run, and compare models directly on your device.

Firethering