All tags
Company: "ai21-labs"
Air Street's State of AI 2025 Report
glm-4.6 jamba-1.5 rnd1 claude-code reflection mastra datacurve spellbook kernel figure softbank abb radicalnumerics zhipu-ai ai21-labs anthropic humanoid-robots mixture-of-experts diffusion-models open-weight-models reinforcement-learning benchmarking small-language-models plugin-systems developer-tools agent-stacks adcock_brett achowdhery clementdelangue
Reflection raised $2B to build frontier open-weight models with a focus on safety and evaluation, led by a team with backgrounds from AlphaGo, PaLM, and Gemini. Figure launched its next-gen humanoid robot, Figure 03, emphasizing non-teleoperated capabilities for home and large-scale use. Radical Numerics released RND1, a 30B-parameter sparse MoE diffusion language model with open weights and code to advance diffusion LM research. Zhipu posted strong results with GLM-4.6 on the Design Arena benchmark, while AI21 Labs' Jamba Reasoning 3B leads tiny reasoning models. Anthropic introduced a plugin system for Claude Code to enhance developer tools and agent stacks. The report also highlights SoftBank's acquisition of ABB's robotics unit for $5.4B and the growing ecosystem around open frontier modeling and small-model reasoning.
not much happened today
7m-tiny-recursive-model jamba-reasoning-3b qwen3-omni qwen-image-edit-2509 colbert-nano agentflow samsung lecuun ai21-labs alibaba coreweave weights-biases openpipe stanford recursive-reasoning density-estimation multimodality long-context retrieval serverless-reinforcement-learning agentic-systems model-efficiency reinforcement-learning transformers rasbt jm_alexia jiqizhixin randall_balestr corbtt shawnup _akhaliq
Samsung's 7M Tiny Recursive Model (TRM) achieves superior reasoning on ARC-AGI and Sudoku with fewer layers and MLP replacing self-attention. LeCun's team introduces JEPA-SCORE, enabling density estimation from encoders without retraining. AI21 Labs releases Jamba Reasoning 3B, a fast hybrid SSM-Transformer model supporting up to 64K context tokens. Alibaba's Qwen3 Omni/Omni Realtime offers a unified audio-video-text model with extensive language and speech support, outperforming Gemini 2.0 Flash on BigBench Audio. Alibaba also debuts Qwen Image Edit 2509, a top open-weight multi-image editing model. ColBERT Nano models demonstrate effective retrieval at micro-scale parameter sizes. In reinforcement learning, CoreWeave, Weights & Biases, and OpenPipe launch serverless RL infrastructure reducing costs and speeding training. Stanford's AgentFlow presents an in-the-flow RL system with a 7B backbone outperforming larger models on agentic tasks. This update highlights advances in recursive reasoning, density estimation, multimodal architectures, long-context modeling, retrieval, and serverless reinforcement learning.
not much happened today
grok-4 jamba ernie-4.5 claude-4-sonnet claude-4 kontext-dev ai21-labs hugging-face baidu perplexity-ai deepmind anthropic reinforcement-learning fine-tuning energy-based-transformers ssm-transformer context-windows length-generalization recurrent-neural-networks attention-mechanisms 2-simplicial-attention biomedical-ai instruction-following open-weight-models python-package-management _philschmid corbtt jxmnop sedielem _akhaliq slashml alexiglad clementdelangue _albertgu tri_dao theaitimeline deep-learning-ai
Over the holiday weekend, key AI developments include the upcoming release of Grok 4, Perplexity teasing new projects, and community reactions to Cursor and Dia. Research highlights feature a paper on Reinforcement Learning (RL) improving generalization and reasoning across domains, contrasting with Supervised Fine-Tuning's forgetting issues. Energy-Based Transformers (EBTs) are proposed as a promising alternative to traditional transformers. AI21 Labs updated its Jamba model family with enhanced grounding and instruction following, maintaining a 256K context window. Baidu open-sourced its massive 424 billion parameter Ernie 4.5 model, while Kontext-dev became the top trending model on Hugging Face. Advances in length generalization for recurrent models and the introduction of 2-simplicial attention were noted. In biomedical AI, Biomni, powered by Claude 4 Sonnet, demonstrated superior accuracy and rare disease diagnosis capabilities. Additionally, the Python package manager
uv received praise for improving Python installation workflows. DeepSeek's Open Source Stack
qwen-qwq-32b start character-3 gemini gemini-2.0 mercury-coder gpt-4.5 jamba-mini-1.6 gemini-2.0-flash gpt-4o-mini mistral-small-3 mistral-ocr deepseek pyspur hugging-face togethercompute hedra-labs google-deepmind deeplearningai openai ai21-labs mistral-ai fine-tuning benchmarking multimodality code-generation diffusion-models model-performance model-optimization ocr embedding-models context-windows runtime-limits _akhaliq lmarena_ai reach_vb danielhanchen _philschmid aidan_mclau vikhyatk jerryjliu0
DeepSeek's Open Source Week was summarized by PySpur, highlighting multiple interesting releases. The Qwen QwQ-32B model was fine-tuned into START, excelling in PhD-level science QA and math benchmarks. Character-3, an omnimodal AI video generation model by Hedra Labs and Together AI, enables realistic animated content creation. Google DeepMind introduced the Gemini embedding model with an 8k context window, ranking #1 on MMTEB, alongside the Gemini 2.0 Code Executor supporting Python libraries and auto-fix features. Inception Labs' Mercury Coder is a diffusion-based code generation model offering faster token processing. OpenAI released GPT-4.5, their largest model yet but with less reasoning ability than some competitors. AI21 Labs launched Jamba Mini 1.6, noted for superior output speed compared to Gemini 2.0 Flash, GPT-4o mini, and Mistral Small 3. A new dataset of 1.9M scanned pages was released for OCR benchmarking, with Mistral OCR showing competitive but not top-tier document parsing performance compared to LLM/LVM-powered methods. "Cracked engineers are all you need."
not much happened today
jamba-1.6 mistral-ocr qwq-32b o1 o3-mini instella llama-3-2-3b gemma-2-2b qwen-2-5-3b babel-9b babel-83b gpt-4o claude-3-7-sonnet ai21-labs mistral-ai alibaba openai amd anthropic hugging-face multimodality ocr multilinguality structured-output on-prem-deployment reasoning benchmarking api open-source model-training gpu-optimization prompt-engineering function-calling
AI21 Labs launched Jamba 1.6, touted as the best open model for private enterprise deployment, outperforming Cohere, Mistral, and Llama on benchmarks like Arena Hard. Mistral AI released a state-of-the-art multimodal OCR model with multilingual and structured output capabilities, available for on-prem deployment. Alibaba Qwen introduced QwQ-32B, an open-weight reasoning model with 32B parameters and cost-effective usage, showing competitive benchmark scores. OpenAI released o1 and o3-mini models with advanced API features including streaming and function calling. AMD unveiled Instella, open-source 3B parameter language models trained on AMD Instinct MI300X GPUs, competing with Llama-3.2-3B and others. Alibaba also released Babel, open multilingual LLMs performing comparably to GPT-4o. Anthropic launched Claude 3.7 Sonnet, enhancing reasoning and prompt engineering capabilities.
not much happened today
phi-4 reinforce++ arc-agi-2 ai21-labs ollama langchain togethercompute groq reinforcement-learning ppo model-optimization memory-efficiency python-packages vision text-extraction frontend-code-generation workflow-automation coding-agents compute-cost-reduction ethical-ai agi-benchmarks scam-alerts sebastien-bubeck fchollet tom-doerr arohan_ bindureddy hwchase17 jonathanross321 clementdelangue vikhyatk
Sebastien Bubeck introduced REINFORCE++, enhancing classical REINFORCE with PPO-inspired techniques for 30% faster training. AI21 Labs released Phi-4 under the MIT License, accessible via Ollama. François Chollet announced plans for ARC-AGI-2 and a next-generation AGI benchmark. LangChain launched 10 new integration packages to boost LLM application development. Tom Doerr introduced Ollama-OCR, a Python package for text extraction using vision language models. Arohan optimized Shampoo for memory efficiency, reducing usage from 20 to 6 bytes per parameter. Bindu Reddy showcased CodeLLM's v1 for frontend code generation and highlighted LlamaIndex Workflows for academic summarization and slide generation. Hwchase17 collaborated with Together Compute to enhance WebDev Arena with complex coding agents for LLM coding evaluations. Jonathan Ross detailed Groq's mission to reduce compute costs by 1000x amid rising generative AI spending. Clement Delangue warned about scam alerts involving false claims of association with AI21. Vikhyat K raised concerns about the ethical implications and trade-offs of AGI. Memes and humor included creative AI prompts and critiques of LLM behaviors.
Everybody shipped small things this holiday weekend
gpt-4o-voice gemini claude jamba-1.5 mistral-nemo-minitron-8b xai google anthropic openai cognition ai21-labs nvidia langchain fine-tuning long-context parameter-efficient-fine-tuning latex-rendering real-time-audio virtual-try-on resource-tags low-code ai-agents workspace-organization model-benchmarking dario-amodei scott-wu fchollet svpino
xAI announced the Colossus 100k H100 cluster capable of training an FP8 GPT-4 class model in 4 days. Google introduced Structured Output for Gemini. Anthropic discussed Claude's performance issues possibly due to API prompt modifications. OpenAI enhanced controls for File Search in their Assistants API. Cognition and Anthropic leaders appeared on podcasts. The viral Kwai-Kolors virtual try-on model and the open-source real-time audio conversational model Mini-Omni (similar to gpt-4o-voice) were released. Tutorials on parameter-efficient fine-tuning with LoRA and QLoRA, long-context embedding challenges, and Claude's LaTeX rendering feature were highlighted. AI21 Labs released Jamba 1.5 models with a 256K context window and faster long-context performance. NVIDIA debuted Mistral-Nemo-Minitron-8B on the Open LLM Leaderboard. LangChain introduced resource tags for workspace organization, and a low-code AI app toolkit was shared by svpino. Legal AI agents and financial agent evaluations using LangSmith were also featured.
not much happened this weekend
jamba-1.5 dream-machine-1.5 ideogram-v2 mistral-nemo-minitron-8b mistral-7b llama-3-8b nous-research cursor-ai gdm george-hotz agibot unitree eth-zurich disney uc-san-diego ai21-labs luma-labs ideogram nvidia mistral-ai meta-ai-fair distributed-ai optimizer inter-gpu-communication low-latency-training open-source humanoid-robots robotics physics-based-motion teleoperation multilingual-models long-context text-to-video text-to-image model-performance george-hotz adcock_brett aman
Nous Research announced DisTrO, a new optimizer that drastically reduces inter-GPU communication by 1000x to 10,000x enabling efficient training on slow networks, offering an alternative to GDM's DiLoCo. Cursor AI gained viral attention from an 8-year-old user and announced a new fundraise, with co-host Aman returning to their podcast. George Hotz launched tinybox for sale. In robotics, AGIBOT revealed 5 new humanoid robots with open-source plans, and Unitree showcased its G1 humanoid robot nearing mass production at $16,000. ETH Zurich and Disney developed an AI system for physics-based robot motion generation from text or images. UC San Diego released ACE, an open-source teleoperation system for controlling multiple robots. AI21 Labs unveiled Jamba 1.5, a multilingual model with 256k context length and permissive licensing. Luma Labs released Dream Machine 1.5 for improved text-to-video generation. Ideogram launched v2 of its text-to-image model with near-perfect text generation. Nvidia and Mistral released Mistral-NeMo-Minitron 8B, a small model outperforming Mistral-7B and llama-3-8b on the Open LLM leaderboard.
Nvidia Minitron: LLM Pruning and Distillation updated for Llama 3.1
llama-3-1-8b llama-3-1 jamba-1.5 claude-3 dracarys-70b dracarys-72b mistral-nemo-minitron-8b mistral-7b nvidia meta-ai-fair ai21-labs anthropic hugging-face pruning knowledge-distillation weight-pruning activation-based-pruning width-pruning kl-divergence teacher-correction prompt-optimization multilinguality long-context mixture-of-experts model-fine-tuning
Nvidia and Meta researchers updated their Llama 3 results with a paper demonstrating the effectiveness of combining weight pruning and knowledge distillation to reduce training costs by training only the largest model from scratch and deriving smaller models via pruning and distillation. The process involves teacher correction, activation-based pruning (favoring width pruning), and retraining with distillation using KL Divergence loss, resulting in better-performing models at comparable sizes. However, distillation incurs some accuracy tradeoffs. Additionally, AI21 Labs launched Jamba 1.5, a hybrid SSM-Transformer MoE model with large context windows and multilingual support. Anthropic updated Claude 3 with LaTeX rendering and prompt caching. An open-source coding-focused LLM, Dracarys, was released in 70B and 72B sizes, showing improved coding performance. The Mistral Nemo Minitron 8B model outperforms Llama 3.1 8B and Mistral 7B on the Hugging Face leaderboard, highlighting pruning and distillation benefits. Research on prompt optimization reveals the complexity of prompt search spaces and the surprising effectiveness of simple algorithms like AutoPrompt/GCG.
super quiet day
jamba-1.5 phi-3.5 dracarys llama-3-1-70b llama-3-1 ai21-labs anthropic stanford hugging-face langchain qdrant aws elastic state-space-models long-context benchmarking ai-safety virtual-environments multi-agent-systems resource-management community-engagement model-performance bindu-reddy rohanpaul_ai jackclarksf danhendrycks reach_vb iqdotgraph
AI21 Labs released Jamba 1.5, a scaled-up State Space Model optimized for long context windows with 94B parameters and up to 2.5X faster inference, outperforming models like Llama 3.1 70B on benchmarks. The Phi-3.5 model was praised for its safety and performance, while Dracarys, a new 70B open-source coding model announced by Bindu Reddy, claims superior benchmarks over Llama 3.1 70B. Discussions on California's SB 1047 AI safety legislation involve Stanford and Anthropic, highlighting a balance between precaution and industry growth. Innovations include uv virtual environments for rapid setup, LangChain's LangSmith resource tags for project management, and multi-agent systems in Qdrant enhancing data workflows. Community events like the RAG workshop by AWS, LangChain, and Elastic continue to support AI learning and collaboration. Memes remain a popular way to engage with AI industry culture.
Jamba: Mixture of Architectures dethrones Mixtral
jamba dbrx mixtral animatediff fastsd sdxs512-0.9 b-lora supir ai21-labs databricks together-ai hugging-face midjourney mixture-of-experts model-architecture context-windows model-optimization fine-tuning image-generation video-generation cpu-optimization style-content-separation high-resolution-upscaling
AI21 labs released Jamba, a 52B parameter MoE model with 256K context length and open weights under Apache 2.0 license, optimized for single A100 GPU performance. It features a unique blocks-and-layers architecture combining transformer and MoE layers, competing with models like Mixtral. Meanwhile, Databricks introduced DBRX, a 36B active parameter MoE model trained on 12T tokens, noted as a new standard for open LLMs. In image generation, advancements include Animatediff for video-quality image generation and FastSD CPU v1.0.0 beta 28 enabling ultra-fast image generation on CPUs. Other innovations involve style-content separation using B-LoRA and improvements in high-resolution image upscaling with SUPIR.