Selected

BBaseten Engineering Blog12:51First-partyModel releases

AI score68

Baseten releases decision model joey-1, with P50 latency of 20ms and $0.01 per million tokens

Baseten's Base Labs has released joey-1, a 2B decision model that directly outputs a single-token label and calibrated confidence, with a P50 end-to-end latency of 20ms, about 2.5 times faster than elephant-1 (49ms), at a price of $0.01 per million tokens, one-tenth that of mammoth-1. joey-1 is post-trained from Qwen3.5-2B and shares the same data recipe and API with elephant-1 and mammoth-1, so switching requires changing only one line. The model is now available on Baseten Model APIs, with zero data retention by default, and its weights are open-sourced on Hugging Face; Model APIs supports a 16k context window, while dedicated deployments can unlock a 262k context window and support fine-tuned weights.

Why read thisThe original text provides joey-1's latency, pricing, and the division of labor among the three model tiers, so readers can judge whether decision-making tasks should switch to a smaller model.

HHugging Face12:47First-partyModel releases

AI score60

NVIDIA Releases Nemotron 3.5 Transcribe and Granary v2 Data Pipeline

NVIDIA has released the open-weight model Nemotron 3.5 Transcribe on Hugging Face and ModelScope, integrating transcription, speaker diarization, translation, and audio question answering into a single checkpoint. It supports transcription in 40 languages, translation between 31 languages and English, and speaker labeling in 13 languages, can process recordings up to 1 hour long, and can label up to 8 speakers. The model ranks first among open-weight models on the English leaderboard of the Hugging Face Open ASR Leaderboard with an average WER of 5.33%, and achieves 2.8% AA-WER on Artificial Analysis with approximately 4B activated parameters. The simultaneously released Granary v2 covers 48 languages and 463,062 hours of speech data, and the accompanying NeMo Speech fine-tuning recipe reduced WER on held-out speakers from 22.44% to 12.17% in a medical case.

Why read thisThe original text provides unified weights for multilingual transcription, speaker diarization, and translation, along with public leaderboards and fine-tuning recipes, making it easy to assess deployment approaches.

MMistral AI02:02First-partyModel releases

AI score87

Mistral Releases Open Weights for Mistral Large 4 (Le Chonk)

Mistral has released the open weights for Mistral Large 4 (Le Chonk), a 1-trillion-parameter natively multimodal model that continues reinforcement learning on top of the preview version, with improved coding and defensive cybersecurity capabilities. It scores 19.7% on FrontierCode 2.0 and 48% on the Artificial Analysis Coding Agent Index, roughly on par with GLM-5.3, but uses an average of 47% fewer tokens per task; on Terminal-Bench 4.0 it improves by 13 percentage points to 41%. The model supports three reasoning modes, uses a modified MIT-based Mistral Ecosystem Open-Weight license, and is priced via API at $0.70 per million input tokens, $0.07 for cached input, and $2.10 for output.

Why read thisThe original text provides open weights, three reasoning modes, and comparisons across multiple benchmarks, allowing readers to judge the trade-offs for self-hosted coding and security scenarios.

MMicrosoft AI Blog09:24First-partyModel releases

AI score67

Microsoft Releases Image Editing Model MAI Image 2.7, Ranked 3rd on Arena Edit

Microsoft AI has released the image editing model MAI Image 2.7, launched in preview form, ranking 3rd on Arena Edit and 2nd on the Artificial Analysis image editing leaderboard, making Microsoft AI the 2nd-ranked lab. The company states that the model outperforms 2.6 on all 10 use cases and 7 categories of editing actions evaluated by Artificial Analysis, leading the next-best model by 24 Elo in scenes and styles. It is currently available for trial on Arena, Artificial Analysis, and MAI Playground, with text-to-image functionality coming soon.

Why read thisThe official rankings of this image editing model on Arena Edit and Artificial Analysis leaderboards are given, along with the improvement margin of use cases relative to 2.6.

HHugging Face09:23First-partyModel releases

AI score83

Liquid AI Releases the LFM3 Series of Models, Including 1.7B Text, Vision, Audio, and Omni Versions

Liquid AI has released the LFM3 series of open-weight models, all branched from LFM3-1.7B-Base, covering text, vision, audio, Omni, and the d1 decision model. LFM3-1.7B-Base adopts a three-hybrid-layer architecture, is pretrained on 22T tokens, and in the mid-training stage over the final 3.5T tokens incorporates more agent trajectories and tool-calling data, with the context window extended to 256K tokens. The company says LFM3-1.7B is about 25% faster than LFM2-1.2B at an 8K context, and scores 38.42 versus 3.02 on WebArena-Verified; LFM3-VL-2B's average score across 62 vision benchmarks rises from 57.6 to 71.4. All models are available for download on Hugging Face, along with a single-GPU fine-tuning notebook.

Why read thisThe original article presents the architectural changes shared by the six models on a common backbone and on-device benchmark data, which can be used to assess the trade-offs of the small-model specialization approach on long-context and agentic tasks.

GGoogle Blog09:22First-partyModel releases

AI score78

Google DeepMind releases world model Genie 4, supporting real-time interactive simulation and native audio

Google DeepMind releases the world model Genie 4, focused on real-time interactive simulation and introducing native audio for the first time in the series. The model is built on capabilities from the Gemini family: Gemini Omni is responsible for making simulations match the appearance and behavior of the real world, while other Gemini models provide the complex reasoning needed to maintain simulation coherence. Users can control scenes in real time using natural language, such as moving the camera, changing the weather, or introducing new characters, and each world also generates ambient sound, music, and interactive dialogue on the fly.

Why read thisThe original text explains that Genie 4 adds native audio and natural language control to real-time interactive simulation, allowing readers to understand the new capability boundaries of world models.

GGoogle Blog08:00First-partyModel releases

AI score82

Google DeepMind Releases Gemini 4 Carbon Frontier Model

Google DeepMind has released Gemini 4 Carbon, calling it its strongest model, ranking first with 59 points on the Artificial Analysis Intelligence Index v5.0 and leading on SEC-bench Pro (88.2%), ExploitGym (54.8%), and CWE-bench (71.0%). The model targets long-horizon engineering tasks; an internal team once had it run continuously for several days to investigate and fix hidden defects in a large codebase, and it rose 89 elo to 4th place (1768) on WebDev Arena. Starting this week, Carbon will first be opened to some cybersecurity defenders, then gradually expanded through the Fairwind Program, and only afterward made available to developers, enterprises, and consumers; after the introductory period, pricing will be $4 per 1M input tokens and $20 per 1M output tokens.

Why read thisThe original text provides Carbon's benchmark results on cybersecurity and long-horizon coding tasks and its phased rollout path, which can be used to assess the pace at which frontier models will be deployed in attack and defense scenarios.

BBaidu12:59First-partyModel releases

AI score62

Baidu Open-Sources Unlimited OCR 2 Document Parsing Model

Baidu has open-sourced Unlimited OCR 2, which uses a new SR-SWA mechanism to compress the attention interaction between text and vision into a fixed-size working reference, thereby parsing documents of almost any length within a limited context window. Baidu says that as documents grow longer, per-step decoding latency and memory usage remain nearly constant, and that the model achieves SOTA performance on MPDocBench and MarathonDocBench, which cover documents ranging from dozens to hundreds of pages.

Why read thisThe official release presents the SR-SWA mechanism and results on two long-document benchmarks, offering insight into the context and latency trade-offs of long-document OCR.

OOpenRouter03:15First-partyModel releases

AI score68

OpenAI GPT-6.1 Terra Launches on OpenRouter

OpenAI's GPT-6.1 Terra is now available on OpenRouter, matching GPT-6 Astra on Agents' Last Exam at one-tenth the price per token. The model completes tasks on average more than 3 times faster than Astra and GPT-6.1 Sol, with pricing at $1/M input and $5/M output, and a context window of 1.05M.

Why read thisOpenRouter launches GPT-6.1 Terra, offering key comparisons: the same score as Astra, one-tenth the price, and three times the speed.

CChatGPT02:13First-partyModel releases

AI score84

OpenAI Releases GPT-6.1 Terra Model

OpenAI has released a new model, GPT-6.1 Terra, which focuses on speed and efficiency, strengthens visual reasoning and precise computer operation, and writes more clearly and naturally. Starting today, the model is being rolled out to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex; the company says it can cover work scenarios ranging from interpreting technical drawings and drafting proposals to filling out forms and editing videos.

Why read thisThe official announcement sets out the new model's positioning in visual reasoning and computer operation, and explains the scope of access for paid users.

QQbitAI16:31Model releases

AI score79

RockAI Technology, together with HKUST and CMU, open-sources EdgeRock-4B, reducing KV Cache by 99.1%

Yanxin Matrix, under RockAI Technology, together with universities including HKUST and CMU, has released and open-sourced EdgeRock-4B. At a 256K context, its KV Cache is fixed at 72MiB, a 99.1% reduction compared with the source model Qwen3.5-4B. A single 80GB GPU can run 512 concurrent long-inference streams, with an output throughput of 12,124 tokens/s, 6.42 times that of the source model; the average pass@1 on AIME 2024–2026 is 82.9%. The model is derived from Qwen3.5-4B through staged conversion, requiring no pretraining from scratch; the weights and inference code have been open-sourced and are available for commercial use, supporting text and image inference.

Why read thisThe original text provides measured comparisons of fixed memory and throughput, based on which readers can assess changes in memory cost for long-inference scenarios.

HHacker News Trending05:42Model releases

AI score80

Dream Labs Releases Robot Model Dream-1 and Opens Public API

Dream Labs has released the world-action model Dream-1, which can control unseen dual-arm parallel-jaw gripper robots to perform unseen tasks via a public API, without fine-tuning or demonstration data from that robot. The company also announced it is emerging from stealth with a $90 million seed round. Dream-1 costs $3 per hour to run, and Dream-1-flash costs $2 per hour; each call returns 1.6 seconds of actions. If zero-shot performance is insufficient, it can be fine-tuned via FaaS with 5 minutes of robot data.

Why read thisDream Labs has disclosed Dream-1's zero-shot performance on unseen robots and its hourly pricing, which can be used to gauge the barrier to deployment for robot foundation models.

UUnite.AI08:46Model releases

AI score78

Block releases full-duplex speech model Dessa, trained from scratch using only human conversation data

Block's internal AI research lab released the full-duplex speech-to-speech model Dessa on October 23, stating that it was trained entirely from scratch on human conversational speech, can listen and speak simultaneously, and natively perceives emotion, sarcasm, and laughter. The model's inference speed is faster than real time, and it can run locally on a MacBook Pro M4 Max or higher configuration; the team completed training with a single NVIDIA GB200 NVL72. Block has published demo recordings and a public waitlist, and has opened early access to a small number of AI safety researchers.

Why read thisDessa was trained from scratch using only speech data, demonstrating how full-duplex conversation handles emotion and interruption timing, and also providing local running requirements.

AArena08:23Model releases

AI score77

Kimi K3.1 Pro released, rises to 8th on Code Arena WebDev

Moonshot AI released Kimi K3.1 Pro, which ranks 8th on Code Arena WebDev with 1,692 points, matching Claude Opus-5 (Max) while costing 70% less, and rising 9 places from Kimi K3 (Max)'s 17th. The model is priced at $2/$10 per million input/output tokens, with a median cost of $1.23 per task, 29% lower than Kimi K3, and its weights will be opened on November 4. Kimi officially calls it a 3T-class open-weight model, featuring a 1M-token context, a 2.8T total parameter/104B active parameter MoE multimodal architecture, and a 65% reduction in hallucination rate.

Why read thisKimi K3.1 Pro's Code Arena ranking and cost data, directly comparable with Claude Opus-5 and the previous-generation K3.

TThe Elec08:23Model releases

AI score82

Moonshot AI Releases Kimi K3.1 Pro, AA Intelligence Index Score of 55 Surpasses GPT-6 Astra

Moonshot AI has released Kimi K3.1 Pro, an upgraded version of Kimi K3, featuring a MoE architecture with 2.8 trillion total parameters and 104 billion activated parameters, activating 16 of 896 experts at a time. According to Artificial Analysis benchmarks, the model ranks fourth on the Intelligence Index with a score of 55, higher than GPT-6 Astra's 53, while the previous generation Kimi K3 scored 44. It ranks first on Terminal Bench 4.0 with 67%, compared with 13% for the previous generation and 64% for Claude Sonnet 5.5; pricing is $2 per million input tokens and $10 per million output tokens, but its output speed of 73 tokens/s is below the average of 81 tokens/s.

NNathan Lambert08:23Model releases

AI score85

Kimi K3.1 Pro Released, Intelligence Index Score of 55 Ranks Fourth

Kimi K3.1 Pro scored 55 on the Artificial Analysis Intelligence Index, ranking fourth, an 11-point improvement over Kimi K3's top score of 44. The model shows notable gains in agentic capabilities, scoring 1,846 Elo on GDPval-AA, 312 Elo higher than K3, and 67% on Terminal-Bench 4.0, above K3's 13%. The model has a 1M-token context window, 2.8T total parameters and 104B activated parameters, supports text and image inputs, and is priced at $2, $10, and $0.20 per 1M input, output, and cache-hit tokens respectively, 33% lower than K3; the weights have not yet been released, and Moonshot expects to release them shortly after launch.

UUnsloth08:18First-partyModel releases

AI score72

NVIDIA releases Nemotron-3.5-Bolt-E4B, which can run locally on 6GB of memory

NVIDIA releases Nemotron-3.5-Bolt-E4B, which can run locally on devices with 6GB of memory, targeting on-device devices such as phones. The model is trained based on Gemma 4, supports vision, and Unsloth also provides GGUF files and a guide link. NVIDIA says it targets scenarios such as local assistants, interactive characters, and robots that can converse and call tools.

Why read thisThe original text gives the model's positioning of running locally on 6GB of memory and its visual capabilities, so readers can judge the usable boundaries of small models in on-device assistant scenarios.

MMoonshot AI · Kimi Blog08:13First-partyModel releases

AI score87

Moonshot AI releases Kimi K3.1 Pro: a 3T-class open-weight model with a 1 million token context

Moonshot AI releases Kimi K3.1 Pro, the first model in the K3.1 series and its strongest 3T-class open-weight model to date, with a 1 million token context window, alongside Kimi K3.1 Pro Highspeed, which is 2-3 times faster. The model is now available on Kimi.com, Kimi Work, Kimi Code and the Kimi API, with full model weights to be released on November 4. On Terminal-Bench 4.0, K3.1 Pro scores 65.7%, higher than GPT-6 Astra's 58.2% and slightly below Claude Opus 5.5's 66.4%; on the internal benchmark Kimi Code Bench 3.0 Long-horizon it scores 61.2%, higher than K3's 55.95%.

Why read thisThe official disclosure covered K3.1 Pro's benchmark performance on coding, 3D and financial tasks, as well as the trade-offs made on cybersecurity data before open-sourcing the weights.

Ffal23:17Model releases

AI score70

Kling 4.0 launches on fal and opens its API

Kling 4.0 has publicly launched on the fal platform, with its API open to all customers. The model supports native video generation with audio up to 30 seconds, up to 15 image/video/element references, and precise control of up to 10 keyframes, and improves motion, character consistency, and cinematic storytelling.

Why read thisKling 4.0 launches on fal and opens its API; the original text lists specific capability changes such as duration, reference materials, and keyframe control.

UUnite.AI09:42Model releases

AI score79

Canva Releases Its In-House Image Model Canva Craft

On October 22, Canva released Canva Craft, the first foundation model trained from scratch by its internal lab Canva Research. It is now available to Canva users worldwide through Dream Lab and to developers via a public API. The model uses a two-stage architecture: first a planning model parses the prompt and generates a composition plan, then a rendering model outputs images at up to 2K resolution supporting more than 30 aspect ratios, and can generate text in 55 languages. Canva says Craft debuted fourth on DesignArena, and previews that two faster variants, Craft-fast and Craft-ultrafast, are coming soon.

Why read thisThe original article breaks down Craft's two-stage architecture of first planning the composition and then rendering, and how it serves editable layers, making it easier to understand the differences of design-oriented image models.