Live intelligence

Latest AI updates

77 updates

Models

Baidu Open-Sources Unlimited OCR 2 for Hundred-Page Document Parsing

Baidu released Unlimited OCR 2, an open-source 3.34B vision-language model for parsing very long documents like books and filings in a single request. It introduces SR-SWA, a sparse attention mechanism that keeps per-step decoding costs nearly constant as document length grows. The model achieves 77.1 on MarathonDocBench, far ahead of competitors, though all benchmark figures are self-reported. Weights are available under the MIT license with code on GitHub.

MarkTechPostOct 27, 2026Score 75

BaiduUnlimited OCR 2SR-SWA

Models Editor selected

OpenAI introduces GPT-6.1 Terra, a midrange model focused on visual reasoning and computer use

OpenAI released GPT-6.1 Terra, a fast, efficient model that emphasizes detailed visual reasoning, precise computer use, and clearer natural writing. It is positioned as a midrange option between the more capable Sol and the cheaper Luna, with pricing at $1 per million input tokens and $5 per million output tokens. The model is described as a significant step up from GPT-5.6 Terra and particularly suited for tasks such as drafting proposals, interpreting technical drawings, filling forms, and edi

Why it matters: This provides concrete details about a new OpenAI model release, including positioning, pricing, and intended use cases, which is directly useful for AI practitioners and subscribers.

Simon WillisonOct 27, 2026Score 80

OpenAIGPT-6.1computer use

Models

Nanbeige Releases Nanbeige4.3-33B, an Apache 2.0 Agentic Coding Model

Nanbeige LLM Lab, part of the recruitment platform BOSS Zhipin, released Nanbeige4.3-33B on Hugging Face under Apache 2.0. It is a 33B dense text model with a 524,288-token context window, aimed at agentic coding and office-style 'cowork' tasks, with a separately published Base model for custom post-training. Nanbeige reports 80.9 on Terminal-Bench 2.1 and 65.5 on DeepSWE against Qwen3.8-27B, Muse-Glimmer-30B and Gemma4-31B, though several competitor figures in its table conflict with those labs

MarkTechPostOct 26, 2026Score 72

Nanbeigeagentic coding modelsopen-weight LLMs

Models Editor selected

Hugging Face Releases Carbon-X, a 14.8B Open DNA Language Model

Hugging Face Bio has released Carbon-X, an Apache 2.0 decoder-only model with 14.8B parameters that processes raw DNA and English in a single token stream and answers in natural language. Built on Qwen3-14B with a hybrid tokenizer adding 5,515 DNA tokens over 6-mers, it was trained in four stages and runs on one 80GB GPU or quantized GGUF builds. Hugging Face reports a 66.7% average across five DNA benchmarks, leading on BioReason Coding at 74.9% but weaker on GenomeQA splice-site questions at 2

Why it matters: It gives readers a concrete, openly licensed DNA model with weight-level access, deployment guidance and self-reported benchmarks for judging practical use.

MarkTechPostOct 26, 2026Score 80

Carbon-XHugging Facegenomics

Models Editor selected

NVIDIA Releases Nemotron 3.5 Bolt E4B, an Open Small Model for Local Agents

NVIDIA released Nemotron 3.5 Bolt E4B, an open small language model built by RL post-training of Google's Gemma 4 E4B without supervised fine-tuning. It has 8B total and 4B effective parameters with 64K context, targeting local voice assistants, game NPCs and robot interfaces on RTX, Jetson and Spark hardware. NVIDIA reports top scores on 15 of 21 benchmarks in its own harness, but Qwen3.5-4B leads on Tau2-Bench and long-context tests. Weights, data and training recipes are public, with the Q4_K

Why it matters: It gives developers a concrete open-weight option for running capable tool-calling agents locally on constrained hardware, with disclosed training methods and honest benchmark weaknesses.

MarkTechPostOct 24, 2026Score 83

NVIDIANemotronGemma

Models

Moonshot AI Releases Kimi K3.1 Pro: 2.8T-Parameter Open-Weight MoE Model with 1M-Token Context

Moonshot AI has released Kimi K3.1 Pro, a post-trained successor to Kimi K3 with 2.8T total and 104B active parameters in a mixture-of-experts architecture and a 1M-token context window. It is available now through an OpenAI-compatible API ($2/$10 per million input/output tokens, cheaper than K3), with a 2-3x faster Highspeed variant. Moonshot reports near-frontier coding results, including 51.7 on SWE-Marathon and 65.7% on Terminal-Bench 4.0, but a large gap on science terminal tasks (38.1 on T

MarkTechPostOct 24, 2026Score 75

Moonshot AIKimi K3.1 ProLLM releases

Models Editor selected

Moonshot releases Kimi K3.1 Pro with open weights promised by Nov 4

Moonshot launched Kimi K3.1 Pro on its API, priced at $2.00 per million input tokens and $10 per million output tokens, with a higher-priced "Highspeed" variant that runs 2-3x faster. The model is not yet open weights, but Moonshot promises full weights by Nov 4, 2026. Moonshot's own benchmarks place it near Claude Opus 5.5 and GPT-6 Astra, and Artificial Analysis scored it 55, behind only Opus 5.5 at 58. Simon Willison tested it through the Moonshot API, including a pelican SVG comparison acros

Why it matters: It signals a soon-to-be open-weight model reaching near-frontier benchmark standing at competitive API pricing, which matters to teams choosing models.

Simon WillisonOct 24, 2026Score 80

MoonshotKimi K3.1 Proopen-weight models

Models Editor selected

Gradium Open-Sources Phonon, a 31M-Parameter On-Device TTS Model Under Apache 2.0

Gradium, a Paris voice AI company spun out of Kyutai, has released Phonon, a family of small text-to-speech models that run fully on-device. The English Phonon 31m ships under Apache 2.0 at 64.5 MB (q8 GGUF) with 31M parameters and 37 built-in voices, producing 24 kHz mono audio with no cloud call or GPU. It uses a causal transformer backbone, SentencePiece tokenization without phoneme conversion, a single-step flow-matching head and a reduced Mimi decoder. Gradium's own evaluations report first

Why it matters: Provides concrete deployment, licensing and performance details for a new openly licensed on-device TTS model developers can self-host.

MarkTechPostOct 23, 2026Score 81

Gradium Phononon-device TTSopen-weight speech models

Models Editor selected

IBM Releases Granite Speech 5.0 2B, an Apache 2.0 Multilingual Speech Model

IBM released Granite Speech 5.0 2B, a 2B-parameter open speech-language model for transcription and translation in 6 languages under Apache 2.0. It combines a 16-block Conformer speech encoder with a window q-former and a fine-tuned granite-4.0-1b-base backbone, and adds language-ID prompting, keyword biasing up to 2,000 terms, and requested punctuation and capitalization. IBM reports 5.68% average English WER on the Open ASR Leaderboard, ranking second among open-weight models behind Qwen3-ASR-

Why it matters: Developers get a concrete, benchmarked open-weight ASR and translation release with deployment options and licensing details they can evaluate directly.

MarkTechPostOct 22, 2026Score 83

IBM Granitespeech recognitionopen-weight models

Models Editor selected

Recursive Releases Flint, an Open-Weight 40B MoE Local Coding Model

Recursive Superintelligence released Flint, a text-only mixture-of-experts coding model with 39.8B total and roughly 2.5B active parameters, under Apache 2.0 on Hugging Face. It targets coding agents, supports reasoning and OpenAI-compatible tool calls, and runs in about 23 GB at 4-bit on 48 GB Apple Silicon Macs or NVIDIA GPUs. Recursive reports 74.2% on SWE-bench Verified and 54.9% on Terminal-Bench 2.1, trailing the strongest dense models on the latter. The company says its Eureka Machine aut

Why it matters: It gives readers concrete details on a new open-weight local coding model, including architecture, hardware needs and self-reported benchmark results.

MarkTechPostOct 22, 2026Score 83

Recursive SuperintelligenceFlintopen-weight coding models

Models

NVIDIA Expands Open Model Portfolio Across Language, Robotics, Biology and Quantum

NVIDIA announced a broad expansion of its open model family, including Nemotron 3.5 Super VL and Bolt language models, Nemotron Retriever embedding and reranking models, and the new Kumo foundation models for structured business data. The lineup also covers Cosmos 3.5 world models, Alpamayo 2 Edge for autonomous vehicles, BioNeMo RFD4-Proteina for protein design and Ising for quantum computing. The company cites European partners including Mistral, Black Forest Labs and H-Company, plus enterpris

NVIDIA BlogOct 21, 2026Score 79

NVIDIAOpen-weight modelsEnterprise AI

Models

OpenAI says GPT-6 models improve on legacy programming languages like COBOL and Fortran

OpenAI improved its latest models for low-resource programming languages including COBOL, Fortran and Visual Basic. On the new MultiLangBench, GPT-6 Astra reportedly reaches an 87.4 percent test pass rate across 20 languages, up from about 38 percent for GPT-5.6 Terra. For COBOL migration, GPT-6.1 Sol is said to reach up to 97.9 percent accuracy, which could matter for modernizing critical financial infrastructure.

The DecoderOct 21, 2026Score 73

OpenAIGPT-6legacy programming languages

Models

Inception Releases Mercury 3 Diffusion LLM at 1,068 Tokens per Second

Inception has launched Mercury 3, a diffusion language model that refines answers in parallel rather than token by token, claiming 1,068 tokens per second on NVIDIA GPUs and a 1M-token context window. Inception's own benchmark table shows Mercury 3 leading GPT-6 Luna and Claude Haiku 5.5 on 8 of 12 tests, with its biggest gains on SWE-Bench Atlas (51.08) and IFBench (76.84), while Haiku 5.5 still wins DeepSWE v1 and MRCR v2. The API-only model is available through the Inception API Platform, Ope

MarkTechPostOct 20, 2026Score 73

InceptionMercury 3diffusion language models

Models

Nace AI releases Drex Max, a multimodal decision model for computer-use agents

Nace AI has launched Drex Max, a hosted model that reads text and images and returns calibrated probability distributions over named questions rather than generating tokens, aimed at computer-use and browser agents. Nace reports 67.59 on the public Decision Index 0.3.2, 6.9 points above TypeSafe's Jev 1.14, alongside 98.2 on BFCL and 92.2 on JevBench, though it trails Jev 1.14 on MMLU-Pro and HLE. The API accepts up to eight images and 16 questions per request with a 262,144-token context, at $0

MarkTechPostOct 20, 2026Score 73

Nace AIDrex Maxcomputer-use agents

Models

MiniMax M3.1 Flash matches DeepSeek on coding benchmark at lower cost

MiniMax's M3.1 Flash scores 42 on the Artificial Analysis Intelligence Index, a 13-point gain over its predecessor at half the cost. It also reaches 73.7 percent on the DeepSWE v1.1 coding benchmark for under $0.30 per task, matching DeepSeek V4.1 Flash 1017 on both performance and price. The Decoder frames the release as evidence of an ongoing price competition among Chinese AI model makers.

The DecoderOct 20, 2026Score 73

MiniMaxDeepSeekAI model pricing

Models Editor selected

MiniMax Releases M3.1 Flash: 428B MoE Agentic Model at $0.15/M Input Tokens

MiniMax has launched MiniMax M3.1 Flash on its API, a multimodal mixture-of-experts model with 428B total parameters, about 23B active, and a 1,000,000-token context window. It is priced at a flat $0.15 per 1M input tokens and $0.60 per 1M output tokens, half of M3's standard per-token rate. MiniMax reports near-frontier agentic results, including a top Automation-Bench v1.0.6 score of 62.2, but weaker hard-coding results such as 38.1 on Terminal Bench 4.0. Weights are due October 23, 2026, with

Why it matters: It gives developers concrete pricing, context, benchmark and deployment details for a newly released low-cost agentic model, with vendor-reported figures clearly flagged.

MarkTechPostOct 19, 2026Score 80

MiniMaxMiniMax M3.1 Flashmixture-of-experts

Models Editor selected

Mistral Releases Timestral, a CPU-Friendly Time Series Foundation Model Family

Mistral AI has released Timestral, a family of four zero-shot time series forecasting models ranging from 5M to 209M parameters, with open weights and inference code. The 209M checkpoint is reported by Mistral to rank first among pretrained zero-shot models on GIFT-Eval, TIME and fev-bench, though several margins are narrow. The models support long context, 21 quantiles per step and covariate-aware attention, and run inference on standard server CPUs without a GPU. Weights are released under a M

Why it matters: It gives developers a concrete, openly weighted forecasting model family with CPU-only inference and benchmark comparisons against TimesFM-3, Chronos-2 and Toto 2.0.

MarkTechPostOct 19, 2026Score 83

Mistral AItime series foundation modelsGIFT-Eval

Models Editor selected

Kyutai Releases REKI, a 2.4B Vision-Language-Action Model with Cached Register Memory

Kyutai, with Inria and École des Ponts, has released REKI, a 2.4B-parameter vision-language-action model for robot manipulation that was pretrained only on public data. It pairs a Qwen3.5-2B backbone with a 200M-parameter flow-matching action expert, and adds memory through 32 learned register tokens cached over the last three frames rather than feeding past frames into the VLM. The authors report 90.4% on LIBERO-Plus and 65.8% on RoboTwin Clean2Rand, with plain-language controls cutting arm osc

Why it matters: It offers a concrete, openly released VLA and a cheap memory design that robotics practitioners can inspect, reproduce and finetune.

MarkTechPostOct 18, 2026Score 81

Kyutaivision-language-action modelsrobot manipulation