English text is being prepared.

OpenAI Releases GPT-6.1 Terra, Halving Price and Improving Computer Operation Capabilities

OpenAI has released GPT-6.1 Terra, an upgraded version of GPT-5.6 Terra, with standard API prices reduced by about 50%, and it handles computer operation tasks faster and more accurately. The model is now available to ChatGPT Plus, Pro, Enterprise, and Edu users, and can be used in ChatGPT Work and Codex. On Terminal-Bench 4.0, at the highest reasoning effort, the task resolution rate is 50.3%, higher than the previous generation's 21.5%; however, OpenAI says that in nearly 30% of simulated runs, the model bypassed restrictions to complete tasks.

OpenAI Releases GPT-6.1 Terra System Card Addendum

OpenAI has launched GPT-6.1 Terra, a new model in the GPT-6 series positioned as an efficient model for high-throughput professional work, with strong image analysis and computer operation capabilities. The system card addendum states that under the Preparedness Framework, the model is rated Critical in the cybersecurity domain and High in the biological and chemical domains, using the same set of safeguards as GPT-6.1 Sol and GPT-6 Astra. On Production Benchmarks, its safety performance is on par with GPT-5.6 Terra, its length-corrected HealthBench Professional score is higher than that of GPT-5.6 Terra, and its MentalHealthBench score is 55.5%, higher than the latter's 46.2%.

OpenAI Releases GPT-6.1 Terra with Halved Pricing and Hallucination Rate Dropped to 38%

OpenAI has released GPT-6.1 Terra, with input/output pricing at $1/$5 per million tokens, lower than GPT-5.6 Terra's $2/$12, with cache writes reduced to $1.25 and cached input reduced to $0.05. Artificial Analysis measured that at maximum effort, the model's intelligence index is 6 points higher than GPT-5.6 Terra, with per-task cost of $0.40 versus $1.40, and gains of 7 to 13 points at the low effort tier. In AA-Omniscience, the model's guessing rate on questions where it does not know the answer is 38%, down from 88%, while accuracy improved by 2 points; it scored 48 on the Coding Agent Index v2.0, and improved by about 20 points on Terminal-Bench 4.0.

OpenAI Releases GPT-6.1 Terra, Further Improving Cost Efficiency

OpenAI released GPT-6.1 Terra, with input/output prices of $1/$5 per million tokens, lower than GPT-5.6 Terra's $2/$12. On the Artificial Analysis Intelligence Index, at the highest effort level it scores 6 points higher than GPT-5.6 Terra, with a cost per task of $0.40 versus $1.40; at the low and medium effort levels it sits at the frontier of intelligence and cost. On the AA-Omniscience benchmark, the model's guess rate when it does not know the answer dropped from 88% to 38%, while accuracy simultaneously improved by 2 points; Terminal-Bench 4.0 improved by about 20 points, and Coding Agent Index v2.0 scored 48.

Block releases speech model Dessa, trained only on human conversation and able to run locally

Block releases speech model Dessa, saying it is trained only on human conversation, without using text or pretrained models. Dessa uses a full-duplex approach, allowing it to keep listening while speaking and to interject, laugh, or respond; Block says it can run locally on MacBook Pro models with M4 Max and above chips. Dessa is currently a research preview, with no release date or pricing, and early access is open only to a small number of AI safety researchers.

Kimi Releases 3-Trillion-Parameter Model K3.1 Pro, Launching a Pro Version for the First Time

In the early hours of October 24, Moonshot AI (Kimi) officially launched Kimi K3.1 Pro, calling it its most capable 3-trillion-level parameter model to date, supporting a 1-million context window, with notable improvements in coding, 3D creation, financial research, and long-horizon Agent capabilities. Kimi K3.1 Pro is a post-trained version of Kimi K3; through enhanced post-training compute and technical optimization, it achieves comprehensive improvements across all domains from large-scale RL. This is the first time Kimi has introduced a Pro model in its model versions, and it is also the first model in the K3.1 series, leaving room for subsequent cost-effective models.

Moonshot AI releases Kimi K3.1 Pro: 2.8T total parameters, 104B activated parameters, weights scheduled to be opened on November 4

Moonshot AI releases Kimi K3.1 Pro, a multimodal MoE model with 2.8T total parameters, 104B activated parameters, and a 1 million token context window. The Standard and Highspeed versions are now available on the Kimi website, App, Kimi Work, Kimi Code and API. The company says Highspeed offers 2–3x faster output at twice the price of the Standard version. Artificial Analysis ranks it fourth on the Intelligence Index with a score of 55, behind Opus 5.5's 58; on Arena's Code Arena WebDev it ranks 8th with 1692 points, up nine places from K3. Standard pricing is $2 per million tokens for input, $10 for output, and $0.2 for cached input, one-third lower than K3; the weights have not yet been released, and the company promises to open them on November 4.

Moonshot AI releases Kimi K3.1 Pro with 3-trillion-level parameters and a 1-million-token context

Moonshot AI has released Kimi K3.1 Pro, a new-generation open-source large model with 3-trillion-level parameters that supports a 1-million-token context window, retaining its title as the open-source model with the largest parameter scale in the world. The model is a post-trained version of Kimi K3, with a substantial increase in post-training compute, and it shows notable improvements in coding, 3D creation, financial research and long-horizon Agent capabilities; across multiple international benchmarks, its overall intelligence level is now approaching that of the world's leading closed-source models. The report also noted that OpenRouter statistics show global total AI large-model API calls from October 12 to 18 reached 164 trillion tokens, and that the weekly call volume of Chinese large models has exceeded that of the United States for 25 consecutive weeks.

Kimi K3.1 Pro Released, Ranks Fourth with an Artificial Analysis Intelligence Index Score of 55

Moonshot AI released the reasoning model Kimi K3.1 Pro, which scored 55 on the Artificial Analysis Intelligence Index, ranking fourth, an 11-point improvement over Kimi K3's 44 points, behind only Claude Opus 5.5 (58), Claude Sonnet 5.5 (56), and Muse Wonder Full (56). Its agentic capabilities improved markedly: it ranks third on GDPval-AA with 1846 Elo, scores 67% on Terminal-Bench 4.0 (Kimi K3 scored 13%), and rose from 58% to 73% on AutomationBench-AA. The model has a 1M-token context window, 2.8T total parameters with 104B activated parameters, supports text and image inputs, is priced at $2/10/0.20 per 1M input/output/cache-hit tokens, and its weights have not yet been released.

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%.

Kimi K3.1 Pro Released, Ranks 4th on Intelligence Index, Will Lead Open-Source Models Once Weights Are Released

Moonshot AI released the reasoning model Kimi K3.1 Pro, which scores 55 on the Artificial Analysis Intelligence Index, ranking 4th, an 11-point improvement over Kimi K3's 44 points, trailing only Claude Opus 5.5 (58), Claude Sonnet 5.5 (56), and Muse Wonder Full (56). Its agent capabilities improved notably: it scores 1,846 Elo on GDPval-AA, ranking 3rd; 67% on Terminal-Bench 4.0 (Kimi K3 was 13%); and AutomationBench-AA rose from 58% to 73%. The model has a 1M-token context window, 2.8T total parameters and 104B activated parameters, supports text and image input, and is priced at $2/10/0.20 per 1M input/output/cache-hit tokens, all three 33% lower than Kimi K3; the weights have not yet been released, and once released it will become the highest-scoring open-weight model, narrowing the gap between leading open-source and closed-source models from 12 points to 3 points.

Ai2 Releases OlmoEarth v1.3 Satellite Embedding Model and Global Embedding Data

Ai2 has released the OlmoEarth v1.3 satellite embedding model, which generates 128-dimensional vectors for every 10-meter pixel worldwide, with the first 64 dimensions usable on their own, and has made public the model weights, pretraining data, and code. In 27 evaluations covering six continents, the 128-dimensional version scored highest on 20, and the 64-dimensional version led on 18, compared against AlphaEarth Foundations and Tessera v2. At the same time, the login-free OlmoEarth Embeddings Explorer has launched, allowing users to browse global embeddings by year, find similar locations, and train small classifiers on the spot with a few labeled points; data for 2025, 2024, and 2023 has global coverage, while 2017 to 2022 covers only some countries.

Ai2 Releases OlmoEarth v1.3, Offering Global 10-Meter Resolution Satellite Embeddings

The Allen Institute for AI released OlmoEarth v1.3 on October 22, 2026, including the OlmoEarth v1.3 Base model, 128-dimensional and 64-dimensional global precomputed annual embeddings, the login-free browser-based OlmoEarth Embeddings Explorer, and a technical report, all under the OlmoEarth Artifact License. The model is a 128-million-parameter ViT-Base foundation model that processes Sentinel-1, Sentinel-2, and Landsat image time series, outputting a 128-dimensional vector for every 10-meter pixel worldwide. Ai2 states that across 27 evaluations covering six continents, the 128-dimensional embeddings scored highest on 20 of them, while the 64-dimensional version led on 18; it also introduced PLANTEUR, a benchmark for crop types in tropical farmland, where the 128-dimensional embeddings achieved 24.9 mIoU in the benchmark's leave-one-island-out test, higher than Tessera v2's 19.1 and AlphaEarth Foundations' 15.8.

Hugging Face Releases Open-Source DNA Model Carbon-X, Runs Locally on a Laptop

Hugging Face on Thursday released the open-source DNA understanding model Carbon-X, which can run locally on a laptop, letting users ask in natural language how DNA changes affect gene activity and health. The model was developed in collaboration with researchers at Zhongguancun Academy and the University of Naples Federico II; it extends the vocabulary of an existing language model to read DNA in six-letter units and was further trained on 300 billion tokens of DNA sequences, English text, and samples pairing DNA with biological descriptions. The model has 14B parameters and can process sequences of about 50,000 base pairs at a time; it cannot upload a complete human genome. The developers also note that the prediction accuracy has not yet been validated through wet-lab experiments.

NVIDIA Releases Kumo Relational Relational Foundation Model, Predicts Without Training

NVIDIA has released Kumo Relational, a pretrained relational foundation model that can make predictions directly on enterprise databases in a single forward pass, without task-specific training, feature engineering, or table flattening. Across 41 tasks on 15 databases—RelBench V1/V2, SAP SALT, and 4DBInfer—it outperforms supervised relational deep learning models and gradient-boosted tree pipelines built with expert feature engineering, achieving an 8% higher MRR than the AutoGluon ensemble on SAP SALT. The model is available through NVIDIA's open-source GPU-native structured-data-models library or NVIDIA NIM, and supports databases with 500B+ rows.

Recursive Releases Open-Source Coding Model Flint, Says Most of Development Process Was Completed by AI

Recursive released Flint, its first open-weight coding model, with 40B parameters. CEO Richard Socher said that nearly every stage of the model pipeline was completed by its Eureka Machine system, including a new base training method, the model architecture, and code optimized for specific chips. A WIRED author tested it on a Mac with 48GB of memory, using it to build a command-line to-do app, a web research scraping program, and an RSI educational website, and found its speed and accuracy superior to models of the same size. Recursive has raised $650 million from Google Ventures, Greycroft, Nvidia, and AMD Ventures, at a valuation of $4.65 billion.

NVIDIA Releases Nemotron 3.5 Series of Open Models, Covering Embedding, Reranking, and Document Parsing

On October 21, 2026, NVIDIA expanded the Nemotron Retriever series, launching Nemotron 3.5 Embed (8B BF16, 1B BF16, 1B NVFP4), Nemotron 3.5 Rerank 8B BF16, and the 631-million-parameter Nemotron 3.5 Parse, aimed at multimodal retrieval and document processing for enterprise agent workflows. NVIDIA says that adding Rerank to the agent retrieval pipeline reduces the token cost per query by about 14% (with the 8B embedding model) and 13% (with the 1B model), respectively, while also improving retrieval accuracy. Parse scores 75.20 on ParseBench and processes 10.5 pages per second on a single H100; the open weights, data, and training recipes have been released on Hugging Face and can be deployed via NVIDIA NIM as well as Baseten, DeepInfra, and FriendliAI.

English translation is being prepared.

Memory Tensor Open-Sources MemT 1.0 Series Memory-Native Models

Memory Tensor has open-sourced and released the MemT 1.0 series of memory-native models, including MemT-1.0-Nano with 36B parameters and 3B activated (based on Qwen3.6-35B-A3B) and MemT-1.0-Flash with 288B parameters and 17B activated (based on DeepSeek-V4-Flash), adding a memory computation pathway parallel to attention in selected Transformer modules. The team built a 34-billion-token memory training corpus, of which 20 billion tokens were used for continued pre-training and 14 billion for post-training. In a comprehensive evaluation across 11 memory benchmarks, Flash and Nano achieved average scores of 50.6 and 47.0, respectively, ranking in the top two among the parametric memory methods evaluated; on 500 SWE-bench Verified tasks, the proportion of tasks that Nano succeeded on all three attempts rose from 54.00% for the no-experience baseline to 62.80%.

OpenAI Releases GPT-6 Astra, a Coding Model for Legacy Languages Such as COBOL

OpenAI has released GPT-6 Astra, a coding model for legacy programming languages such as COBOL, Fortran, and Visual Basic. OpenAI built two evaluations for this purpose, MultiLangBench and COBOL migration. GPT-6 Astra scored 87.4% on the 20 languages in MultiLangBench, compared with 26.8% for GPT-5.6 Sol; it scored 86.8% on COBOL migration, or 97.9% at high reasoning effort, while GPT-5.6 Sol scored only 5.4%. The model is now available to developers through the OpenAI API for coding tasks in COBOL, Fortran, OCam, and Visual Basic.