←
IzwiAI49 posts
IzwiAI avatar

IzwiAI

@IzwiAI

Voice AI Runtime

◷ Joined March 2026▧ 3 Photos and videos
IzwiAI avatar
IzwiAI@IzwiAI·Nov 14

Izwi 0.1.3 adds dictation: hold a shortcut in any app, talk, let go. Your words land at the cursor. Runs on your computer. Audio is never saved. Cleans up your ums. Learns your words. Free and open source: izwiai.com Star on GitHub: github.com/izwi-ai/izwi

0:57
Lennex Zinyando avatar
⟳ IzwiAI retweeted
Lennex Zinyando@zinyando·Oct 29

Liquid AI LFM3 models are now supported in IzwiAI 🚀 Download the desktop application and start playing with LFM3-Audio-2B and LFM3-1.7B. Thanks to the @liquidai team for such great models. Give us a ⭐️ on GitHub - github.com/izwi-ai/izwi Download - izwiai.com/download

Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵

2:17
Lennex Zinyando avatar
⟳ IzwiAI retweeted
Lennex Zinyando@zinyando·Oct 29

Maya-1 and Maya-2 support is officially on @IzwiAI. The frontier voice models for India are now available on your desktop. Download and give us a ⭐️: github.com/izwi-ai/izwi izwiai.com/download

Maya-2 voice model weights are officially live! 🇮🇳🇮🇳 The frontier voice model for India is now completely open-weight. • Outperforms ElevenLabs v3, Sarvam Bulbul v3, and Cartesia Sonic 3.5 • 1.5B+ people covered across all 28 states and 8 union territories • SOTA voice synthesis across 11 official languages • 20 native speakers pre-loaded, ready out of the box • Best-in-class zero-shot voice cloning Built by a team of five from the land of languages. Go build and have fun!

0:50
Liquid AI avatar
⟳ IzwiAI retweeted
Liquid AI@liquidai·Oct 27

Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding > 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages. > 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language. > 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒 Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ 🧵

2:17
Fish Audio avatar
⟳ IzwiAI retweeted
Fish Audio@FishAudio·Oct 27

Introducing Fish Audio Realtime STT (preview): speech-to-text that understands the conversation, not just the words. It knows when to answer, when to keep talking, and when to stop. Ranked #1 of 20 on Cekura’s STT benchmark (2.21% overall word error rate). Built for voice agents. Try it now 👇

1:12
Artificial Analysis avatar
⟳ IzwiAI retweeted
Artificial Analysis@ArtificialAnlys·Oct 23

Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the leading open-weights model once its weights are published @Kimi_Moonshot's new reasoning model gains 11 points over Kimi K3 (max, 44). It trails only Claude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index. Once the weights for K3.1 Pro are released it will be the highest-scoring open weights model and the gap between the leading open weights and proprietary models on the Intelligence Index will narrow from 12 points to 3. Key results: ➤ Large gains in agentic performance: Kimi K3.1 Pro scores 3rd on GDPval-AA with 1,846 Elo, up 312 Elo from Kimi K3, behind only Claude Opus 5.5 (1,866 Elo) and Muse Wonder Full (1,857 Elo). Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0. K3.1 Pro scores 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 (64%). AutomationBench-AA, which tests agentic SaaS workflows, rises from 58% to 73%. ➤ Gains in other evals are more moderate: Humanity's Last Exam rises from 47% to 52%, CritPt from 23% to 25%, and SciCode from 59% to 61%. AA-Omniscience Index rises from 20 to 30, as the hallucination rate falls from 53% to 44% and accuracy rises from 48% to 51%. ➤ Cost per Task roughly equal to Kimi K3 at $1.98 per Task: K3.1 Pro uses 55K output tokens per Intelligence Index task, 14% more than Kimi K3 at 48K, while its first-party token prices are 33% lower. Additional model details: ➤ Context window: 1M tokens. ➤ Size: 2.8T total, 104B active parameters ➤ Input modalities: Text and image ➤ Weights and license: Weights currently not published, Moonshot is expected to release the weights soon after launch ➤ Pricing: $2/$10/$0.20 per 1M input/output/cache hit tokens, 33% lower than Kimi K3 on all three rates. Congratulations to the team at @Kimi_Moonshot on the release!

Post attachment
Lennex Zinyando avatar
⟳ IzwiAI retweeted
Lennex Zinyando@zinyando·Oct 21

Izwi 0.1.0 is stable. Open-source voice AI that runs entirely on your computer. Talk to it in real time, clone and design voices, narrate audiobooks, and transcribe meetings speaker by speaker. Give it a ⭐️ github.com/izwi-ai/izwi Download: izwiai.com

0:49
Mark Zuckerberg avatar
⟳ IzwiAI retweeted
Mark Zuckerberg@finkd·Oct 12

Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.

Post attachment
NVIDIA AI avatar
⟳ IzwiAI retweeted
NVIDIA AI@NVIDIAAI·Sep 23

When several people talk at once, a transcript can get messy fast. Our new Nemotron 3 Diarization model tracks who spoke when, even when voices overlap. It handles up to eight speakers, has 100M parameters, and is now available on @huggingface 🤗

0:35
LightSeek Foundation avatar
⟳ IzwiAI retweeted
LightSeek Foundation@lightseekorg·Aug 25

2026 feels very different from previous years. Single-GPU optimization is becoming increasingly commoditized: profile the timeline, optimize hotspots, adopt TRT-LLM/FlashInfer kernels, have agents write kernels, or borrow ideas across engines. Spec(EAGLE-3/DSpark/DFlash2) and quant(NVFP4 with ModelOpt) are all becoming increasingly standardized. The real differentiation is moving to system-level optimization: PD disaggregation and KV stores — maximizing TPM/GPU under TPS SLAs while pushing cache hit rates higher. Engine optimization is becoming increasingly deterministic and programmatic. But there are still a few months left in 2026. We’ll still see plenty of interesting work ahead — new models, new hardware, and new ideas on existing ones. Looking forward to it!

Lennex Zinyando avatar
⟳ IzwiAI retweeted
Lennex Zinyando@zinyando·Aug 23

That’s what we are doing @IzwiAI, building voice models that understand African languages.

Someone asked what I'd do if I were 17. I'd learn how to build LLMs from scratch, and then train ones as powerful as I could with whatever hardware I could get access to.

Qwen avatar
⟳ IzwiAI retweeted
Qwen@Alibaba_Qwen·Aug 14

We promised open weights for Qwen3.8. Now, time to meet them! 🎉 ⚡ Qwen3.8-27B: - A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows. - 262K native context, easily extendable to 1M tokens via YaRN. - Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0. 🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently. Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now! Download, deploy, and build something we haven't imagined yet. 👀👇 - Hugging Face: huggingface.co/collections/Q… - ModelScope: modelscope.cn/collections/Qw…

Post attachment
Lennex Zinyando avatar
⟳ IzwiAI retweeted
Lennex Zinyando@zinyando·Jul 15

GPT 5.6 Sol is truly amazing. While working on @IzwiAI, a voice AI runtime, Sol completely understands the task and just does it right the first time. The code has gotten a lot more stable in the last week 🚀

Lennex Zinyando avatar
⟳ IzwiAI retweeted
Lennex Zinyando@zinyando·Jul 2

IBM’s Granite Speech 4.1 2B Plus is built for rich transcription: speaker labels, word timestamps, keyword biasing. Best way to test it locally: @IzwiAI. `izwi pull Granite-Speech-4.1-2B-Plus` Local-first voice AI stack. OpenAI-style APIs. github.com/izwi-ai/izwi