Technology

TinyFish published two engineering research articles on agentic search. One defines a 7-phase production architecture with evidence contracts and bounded retries. The other explains why its deep research system dropped a separate planner and per-query subagents.

Google Cloud announced Agent Compute Engine, a compute service that gives every AI agent its own hardware-isolated microVM with no credentials inside. It already runs Gemini agents and enters preview in the coming months.

You.com launched Inference Router, an OpenAI-compatible API for 12 open-weight models with a 99.9% uptime SLA, zero data retention, 100% US hosting, server-side web search and no platform fee.

Liquid AI released RIPL, a real-time inference engine for its LFM models. On one B200, LFM2.5-1.2B on RIPL showed 14.4 to 28.4x lower p95 latency than Qwen3.5-0.8B on vLLM.

Gradium has open-sourced Phonon, a family of on-device text-to-speech models. The English Phonon 31m ships under Apache 2.0, fits in 64.5 MB, and starts speaking in about 7 ms on a laptop CPU.

Cohere Rerank 5 ships in Pro and Fast tiers with a 32K context window, Context Focus passage extraction and calibrated relevance scores. Pro scores 82.2 on BRIGHT and costs $0.05 per 1M tokens.

Explore how to audit and rebuild Datalab's OmniParseBench, test parser ranking certainty with document-level bootstrapping, and evaluate Tesseract and Indra OCR performance on a custom sample.

IBM released Granite Speech 5.0 2B, an Apache 2.0 speech recognition and translation model. It ranks #2 on the Open ASR Leaderboard with 5.68% average English WER, alongside updated Turbo CTC and 4.1 2B Plus models.

Perplexity and NVIDIA introduced JOLT, a post-training method that trains one model as both teacher and student. On MATH it matched GRPO’s accuracy with 6.5x fewer processed training tokens.

Explore a comprehensive coding guide benchmarking Cactus Compute's 16.9 MB speech model against Whisper base. Learn about word error rates, CPU speed, streaming latency, and speech-to-tool integration.

Discover NVIDIA Nsight Python, an open-source interface that allows developers to profile GPU kernels, analyze hardware metrics, evaluate cache and clock controls, and measure kernel fusion directly from Python code.

Recursive has released Flint, an Apache 2.0 local coding model with 40B total and 2.5B active parameters. It reports 74.2% on SWE-bench Verified and 125+ tokens per second on an M4 Pro MacBook Pro.

Vercel released durable-harness, an open-source package that logs every model and tool call so Pi, OpenCode, Claude Code and Codex sessions resume after crashes, timeouts and redeploys on Vercel, Postgres or a laptop.

Mistral AI released Voxtral 3, a family of 3 transcription models, 1 text-to-speech model and a LiveKit-based voice agent SDK. Voxtral Small Transcribe 3 scores 2.1% AA-WER, ranking #3 on Artificial Analysis.

Google, Red Hat and IBM released llm-d 1.0, an open-source Kubernetes-native stack for distributed LLM inference. It adds multi-tenant flow control, zero-downtime upgrades and agent-aware routing across vLLM, SGLang, TPUs and GPUs.

Google and Meta open-sourced TorchTPU, a native PyTorch backend for Cloud TPUs. It runs eager by default, compiles through torch.compile and XLA, and replaces PyTorch/XLA for new workloads.

DeepSeek has open-sourced DeepInfer, the Apache 2.0 inference framework that serves DeepSeek-V4.1-Flash in production. It targets NVIDIA SM100 GPUs and Huawei Ascend 950 NPUs, and ships offline inference only.

Exa 3.0 is a search API upgrade from Exa. It adds a larger embedding model, Exa Lite at $4 per 1,000 requests, and Dynamic Highlights, and leads 8 benchmarks Exa ran itself.

Cloudflare has previewed Clairic, a System One extraction model on Workers AI that returns typed values with confidence scores in one forward pass. It costs $0.10 per million input tokens and charges nothing for output.

Inception released Mercury 3, a diffusion LLM it says runs at 1,068 tokens per second with a 1M-token context. On Inception’s own table it leads GPT-6 Luna and Claude Haiku 5.5 on 8 of 12 benchmarks.

Nace AI released Drex Max, a multimodal decision model that reads text and screenshots and returns calibrated probabilities. Nace reports 67.59 on Decision Index 0.3.2 at $0.06 per 1M input tokens.

MiniMax has released M3.1 Flash, a 428B-parameter MoE model with about 23B active parameters and a 1M-token context. API pricing is $0.15/$0.60 per 1M tokens, with weights due October 23

NVIDIA has released the full Run:ai GPU orchestration platform as open source under Apache 2.0. The release includes KAI Scheduler, CUDA 13.4 GPU fractioning, quota management, and a local Kind dev environment.

Mistral released Timestral, a time series foundation model family from 5M to 209M parameters. The 209M model leads GIFT-Eval, TIME and fev-bench among zero-shot models and runs on CPU.

An AI agent that reads the web pays 2 kinds of cost. The first is the retrieval charge: what a search or fetch API...

Google Research released Harness2, an Apache-2.0 framework that rewrites an agent's harness at test time. One refinement step lifted GLM-5 by 11.5 points on WorkBuddy-Bench with no weight updates.

In this tutorial, we implement Structured Data Models, NVIDIA's new library that puts several pretrained foundation models for tables behind one PyTorch interface, and...

Voxel51 3.0 merges search, curation, annotation and evaluation of multimodal sensor data into one platform for physical AI teams. It adds Voxel51 Agent with 45+ MCP tools, dataset versioning, and a free self-service tier.

Kyutai, with Inria and École des Ponts, has released REKI, a 2.4B vision-language-action model (VLA) for robot manipulation. The REKI vision-language-action model is pretrained...

DeepSeek has just released DeepSeek-V4.1-Flash-1017, a new checkpoint of its V4.1-Flash model. It is live on the DeepSeek website, app and API. DeepSeek claims...

TypeSafe launched Jev on September 15, 2026. It is a decision model, not a chatbot. You define the allowed answers, and Jev returns a...

StepFun has released Step 5 Preview as open weights. FP8 and BF16 checkpoints are live on Hugging Face and ModelScope. Step 5 Preview is...

Perplexity has open-sourced SPACE, the sandbox platform behind every task in Perplexity Computer. The Perplexity SPACE sandbox gives each agent task its own Firecracker...

Liquid AI released d1-ultra, a text and image decision model that returns calibrated probabilities in 1 forward pass. It scores 65 on the Artificial Analysis Decision Index and costs $0.042 per million input tokens.

Meta released Muse Spark Open, an 816B-parameter open-weight MoE model with 42B active parameters and a 1M-token context. It leads open models on MCP-Atlas and runs on 2 M5 Ultra Mac Studios.

NVIDIA released Nemotron 3.5 Super VL, an open 120B-parameter multimodal MoE model with 12B active parameters, 1M-token context and NVFP4 weights that run on 1 DGX Spark, plus 2 Bolt edge models.

Introduction Teams shipping AI agents at scale keep running into the same pattern. Even agents built on top-tier models like Fable and Astra struggle to...

At Open Source Summit Europe 2026, openKylin outlined its vision for transitioning from resource management to intent management in the AI era. The community introduced KART, a unified AI runtime, alongside core desktop features in openKylin 3.0 like the AI Assistant, Global Intelligent Search, and Intelligent Space

TypeSafe AI released Jev 1.14 under the jev-preview alias. The System One model adds experimental image input, a Box bounding-box type and an opt-in 262k-token context, at $0.042 per million input tokens.

SpaceXAI just released Grok Voice 3.0. Grok Voice 3.0 is a SpaceXAI speech-to-speech model that listens, reasons, calls tools and speaks. It now ranks...

Datalab released OmniParseBench, an open OCR benchmark with 16,288 yes/no tests on 2,937 pages in 92 languages. Every score is a pass rate you can slice by tag, script or language.

Fish Audio released Drama 3, a 35B mixture-of-experts text-to-speech model with about 3B active parameters. It takes plain-language direction, covers 83 languages and ships open weights under a revenue-capped community license.

Marktechpost ranks the 9 best AI gateways for 2026. We compare Bifrost, Portkey, LiteLLM, Kong, TrueFoundry, Cloudflare, Vercel, OpenRouter and Orq.ai on latency, pricing, licensing, governance and deployment.

Liquid AI released LFM2.5-24B-A2B, an open-weight 24B Mixture-of-Experts reasoning model with 2B active parameters. It targets local agents, scoring 48.09 on LiveClaw and decoding 184 tokens/s on an Apple M5 Max.

Meta released Muse Wonder and Muse Wonder Full, 1M-context agent models. Muse Wonder costs $4/$20 per million tokens on Meta Model API. Wonder Full, on Muse Code plans, leads DeepSWE and SWE-Marathon.

In this comprehensive coding tutorial, we explore Meta's Rebalancer assignment solver using its C++ core and Python bindings. Learn how to build datacenters, manage task placements across racks, handle constraints and goal priorities, and compare local search against the exact HiGHS solver using published benchmarks.

Z.ai’s GLM-Legal is a legal AI model built on GLM-5.3-Flash. It scores 19.2% all-pass on Harvey’s Legal Agent Benchmark, nearly 3x its base, at $0.15 per test. Access is by application.

Alibaba's Qwen team has introduced QwenGyre, an elastic reinforcement learning RL framework for extreme-long-horizon agent training. QwenGyre is an online reinforcement learning system from...

We checked 12 AI live translation tools, including Palabra.ai, Google Live Translate, DeepL Voice, Teams Interpreter and Wordly, against official pages for languages, latency, platforms and pricing.

TII has released Falcon-ASR, a 1.6B speech recognition model for Arabic with an Emirati focus. It reports 20.92% average WER on 6 Arabic test sets, but weights and an API are not yet available.

Cloudflare released Clef-omni, an Apache-2.0 multimodal decision model built on Qwen3-Omni-30B-A3B. It scores text, images, audio and video in one call for $0.15 per million input tokens, while Clef-flash drops to $0.038.

vLLM Semantic Router released Decision 3.0, 5 Apache-2.0 multimodal decision models from 0.8B to 27B. They answer choice, yes/no and score questions about text and images in one call, without generating text.

OrcaRouter released OrcaCyber Zero 1.5, a gated cybersecurity model with a 1M-token context window. It reports 100% on Cybench and 95.8% on an evaluable CVE-Bench subset, priced at $3.00 / $7.50 per 1M tokens.

Sakana AI’s TMLR paper introduces Multi-Layered Review, a 3-agent Claude-based reviewer, and a 1,164-error Contradiction Benchmark. MLR caught 73.43% of core-claim errors, versus 14.81% for the best prior system.

Microsoft has released Microsoft-Decision-1, a decision model for routing, classification, verification and agent control. Microsoft-Decision-1 is a decision-scoring model that returns a calibrated probability...

Nace.AI has open-sourced Drex 1.5, a 9B decision model that returns a probability for every option in 1 forward pass. It scores 58.08 on Decision Index 0.3.1 and reads up to 128K tokens.

Alibaba's Qwen team has released Qwen-Image-2.1-Turbo, an accelerated checkpoint of its open-weight Qwen-Image-2.1 model. It generates and edits images in 8 denoising steps instead...

OpenAI’s Decisions API is now in public beta on GPT-6 Luna. It returns typed probabilities, choices and scores about 10x faster than the Responses API, billing $0.10 per 1M input tokens with no output charges.

Underdog Saluki 27B is a 7.89 GB, 2-bit GGUF of Qwen3.8-27B under Apache 2.0. It beats the 54 GB original on tool calling but gives up ground on competition math and reasoning.

Google Cloud has introduced the Google Cloud Gemini agent, a single agent for enterprise work. The Gemini agent is a cloud-hosted agent from Google...

JetBrains released Mellum2.1, an Apache 2.0, 12B mixture-of-experts thinking model with 2.5B active parameters. RL in real repositories lifted its SWE-bench Verified score from 2.0 to 47.0.

Architect Financial Technologies has launched Liquid Inference, an LLM router that runs a live auction for every request. Liquid Inference is an LLM inference...

NVIDIA researchers introduced PivotOPD, an on-policy distillation method that trains multi-turn LLM agents to avoid early pivotal mistakes and recover from them, posting the best average against 13 baselines on 3 agent benchmarks.

Perplexity's pplx-embed-v2-late comes in 2 sizes: a 0.6B model built to run on edge devices, and a 9B model for building high-quality indexes. Its best score is 92.4% on MADQA, and its weakest is 61.2% on ViDoRe v3 Markdown. Both are MIT-licensed and ready to self-host.

Unsloth's October 6 security overview explains how Studio checks code, weights, packages and tools before anything runs. Custom model code is scanned and approval is bound to its fingerprint. Flagged weight files are blocked in the load path, package-content findings fail CI, and tools run in probed OS sandboxes. Here is what each checkpoint decides, and what it does not cover.

Anthropic's Claude Haiku 5.5 starts at $0.10 per million input tokens, keeps 1M context, scores 72.4% on OSWorld.

Liquid AI has released Open d1, two open-weight multimodal models in its d1 decision model family. d1-3B reads text and images. d1-omni-600M reads text...

Meta has open-sourced Rebalancer, the C++ and Python library it has used for over 9 years to place shards, servers and traffic. It handles about 40 million assignment problems a day, using local search or MIP solvers like Gurobi, FICO Xpress and HiGHS. It is pip-installable under Apache 2.0.

Explore a comprehensive coding guide to Laya, the open-source zero-shot decision engine. Learn how to implement typed decisions, fit custom temperatures, and build reliable abstention gates using real-world CLINC150 banking data.

Google DeepMind's EmbeddingGemma 2 maps 5 input types into one 768d space and ships today under Apache 2.0.

Mistral AI has released Mistral Large 4, nicknamed Le Chonk, as a public preview. It is a 1.05 trillion parameter Mixture of Experts model with 49 billion active parameters, native image input, and a 1 million token context window, trained on 3,800 NVIDIA Grace Blackwell GPUs in Mistral's own European datacenters. The API is live now; open weights ship end of October 2026.

Reka has released Rho-1, a 19B omni-reasoning model trained from scratch. One network reads and generates text, images, video and robot actions over a shared KV cache. A distilled variant returns a 5.3-second clip in about a second. It is a research preview with no public weights yet.

JEPA-Anything splits a JEPA's single latent target into 4 orthogonal factors, each with its own predictor. Tested across 7 domains, it beat matched JEPA baselines on all 10 dynamics tasks and cut Interventional Pong intervention error by 34.8%.

Together AI has released Together Link, a free, MIT-licensed CLI that connects Claude Code, Codex, OpenCode, Pi, and the Claude and ChatGPT desktop apps to open models like Kimi K3 and GLM 5.3. One install command sets it up, and an Auto router picks a model for each session. Together claims savings of over 50%.

Discover how to construct an end-to-end streaming robotics learning pipeline using the NVIDIA Cosmos3-DROID dataset without local downloads, leveraging byte-range Parquet reads, behavior cloning, and temporal ensembling.

Reflection AI has introduced Beam, its first open-weight model. It is a 501B sparse Mixture-of-Experts model with 23B active parameters, built for coding and agentic work. Reflection says it matches GLM-5.2 on reasoning with 3 to 4x less inference compute. Apache 2.0 weights are due later in October 2026.

One transformer ran candidate generation and ranking in Yandex Music's A/B test without hand-engineered features, lifting likes 11.42%.

Alibaba's Qwen went from an invite-only chatbot in April 2023 to a 2.4-trillion-parameter open-weight model in August 2026. This is the full story, release by release: every major model, its key feature, and how its license changed. Each claim links to its source.

Cantina Security, with Yeta Labs, has released apex-flash-1, an open-weights model trained specifically for vulnerability research. It is a reinforcement learning fine-tune of Z.ai's...

Astra leads computer use, Argon leads legal and finance work, and Sol wins on price for coding agents.

Google Research has unveiled a federated learning system that moves gradient computation from phones into attested server-side TEEs. Access policies are published to Sigstore's Rekor log and the binaries are reproducibly buildable, so central differential privacy can be checked externally. Gboard already uses it for English and Japanese next-word prediction.

Aleph Alpha has released Kolibri, a 78.1B-parameter English-German Mixture-of-Experts model that activates only 3.46B parameters per token. It has a 1M-token context and per-request reasoning effort, and its Apache 2.0 FP8 weights run on a single B200 or H200.

DeepSeek has released official macOS and Windows desktop apps for DeepSeek Harness v0.2, its MIT-licensed agent harness. The preview adds a plugin manager, file and code-change review, and scheduled Automation Tasks. It also supports non-DeepSeek models through OpenAI-compatible endpoints.

TL;DR: Meta's Muse, OpenAI's Dots and Uber's driver assistant share one bet: the agent speaks first. That moves the hard problem from what to...

IBM has made a self-hosted deployment of IBM Bob, its agentic software development platform, generally available. Enterprises can now run Bob on premises, in private or sovereign clouds, and in air-gapped networks. They bring their own model: NVIDIA Nemotron or Poolside Laguna for full isolation, or Claude, Gemini or GPT models through hybrid setups. Optional Premium Packages extend Bob to Java, IBM i and IBM Z modernization.

Prime Intellect has launched Prime Inference, an OpenAI-compatible platform for serving frontier open models on NVIDIA Blackwell. Its GLM-5.3 deployment uses Dynamo, vLLM and NVFP4 KV compression to serve 66 sessions per prefill group at 101 tok/s per user.

Microsoft AI has released MAI-Transcribe-2-Streaming, its first real-time speech-to-text model. It ranks #1 of 38 models on Artificial Analysis AA-WER Streaming. It scores 2.5% WER at 0.13s on final transcripts and 2.5% at 0.12s on first partials. It covers 60 languages with continuous language detection and costs $0.54 per hour during the introductory period. It is available now in public preview on Microsoft Foundry.

Decision AI models answer typed questions with calibrated probabilities instead of generated text. TypeSafe's Jev costs $0.042 per million input tokens and returns responses in 70 to 500 milliseconds. We break down how it works and its benchmarks, then compare it with Fastino's GLiDE, GLiNER2.5-Decide and 4 open-source rivals.

NVIDIA announced a new 64GB configuration of DGX Spark — from Acer, ASUS, Dell, Gigabyte, HP and MSI — its GB10-powered desktop AI system....

Content-based row matching, 6 per-value verdicts and a null rule make OmniExtractBench an extraction benchmark anyone can audit.

AWS's Strands Agents team released Strands Decider 2B, an Apache-2.0 decision model built on Qwen3.5-2B-Base. It returns choices, yes/no probabilities and scores with calibrated confidence in one forward pass, never text. It runs at a 115 ms median on an RTX 3090 and scores 0.723 on the JevBench public set, which makes it a fast local option for routing, tool selection and guardrails in AI agents.

Cloudflare has released Clef (27B) and Clef-flash (9B), open-weight decision models that return typed probabilities instead of text. They are Jev-API compatible, accept images, and run on Workers AI at 209.3 ms and 38.8 ms median latency.

In this comprehensive coding guide, we explore Google Research's Kauldron—a JAX training library optimized for research velocity and modularity. Learn how konfig turns experiments into plain dictionaries, kontext wires components via string paths, and ktyping enforces runtime shape checks.

Cohere has released Embed 5, a new embedding model family. It targets enterprise search, RAG, and agentic retrieval. The model family ships in 2...

NVIDIA has released Kumo Tabular, a new family of tabular foundation models (TFMs) for classification and regression. If you have followed TabPFN or TabICL,...

Perplexity Research and turbopuffer have released pplx-embed-v2-context-9b-preview, a contextual embedding model for RAG pipelines. Each chunk is embedded with the full document in view....

Gemini 4 Argon tops GPT-6 Astra and Claude Opus 5.5 on most benchmarks, but access remains gated today.

OpenAI released GPT-6.1 Sol on September 29, 2026, an upgrade to GPT-6 Sol. It reaches near-Astra results on agentic coding, computer use and professional work at one-fifth of Astra's token prices. It costs $2 input and $10 output per million tokens, and cached input drops to $0.10. It is available now through the OpenAI API, ChatGPT Work and Codex.

Perplexity has released Photon, an in-house retrieval and ranking engine written in Rust. It replaces an open-source engine Perplexity had forked for its AI-native...

Video world models can render convincing clips that still break physics. Butter spreads like paint. Balls pass through walls. A team from NVIDIA, MIT...