Week 43, 2026 · Advanced · for university students

Newly checked in Week 43, 2026

Checked this week: 17 entries, 12 axes. Most: ⑦ Multimodal, 3 entries. Fewer than the past 4-week average of 34 (4 weeks: 26, 18, 71, 20). 10 new actors — Xuyang Liu et al., Anthropic, Government of Singapore, Hanzhang Zhou et al.

17records added

12axes that moved

Oct 19–25dates

1,295total records then

Added, by axis

Axes with the most records come first. A record that spans several axes is counted once, under its main axis.

⑦3 records

Multimodal

Oct 15, 2026Black Forest Labs — Black Forest Labs released the open weights of FLUX.3 Image Klein — one model handles making a picture from text, editing with several reference pictures, and layout control that places regions by writing coordinates into the prompt SourceIt shipped two versions, a 16-step distilled one and a 50-step base one, together with FP8 and NVFP4 quantized weights and a DreamBooth LoRA training script. Only a speed figure was given — the distilled version makes a 1024×1024 picture in about 2 seconds on a single H100 — and no quality scores
Oct 15, 2026Luozheng Qin et al. — MJev-Spatial released — instead of writing out an answer to a spatial question as text, it solves it as a multiple-choice problem, scoring options, numbers and 2D coordinates with projections the model already has and settling the answer in a single computation SourceThe authors reported that, trained on 12.3 million question-answer pairs on top of Qwen3.5-4B, it ranked No. 1 among the 8 models compared, with an overall 65.64%, and was No. 1 on 7 of 12 benchmarks. They reported it was 2.5 times faster than the fastest generative approach and 5.6 times faster than the model it was built on
Oct 14, 2026xAI — xAI released Grok Voice 3.0 — a model that answers speech to speech directly without turning speech into text, offered in the Voice Agent API at $0.08 per minute, with existing grok-voice-latest usage switched over automatically SourceCompany-reported figures: an Artificial Analysis speech-to-speech index of 86.9 (GPT-Live-1 83.2, Gemini 3.8 Live 82.6), Full Duplex Bench 97%, and response latency under 1 second. xAI said it is used in Tesla vehicles and for Starlink customer support, where about 70% of roughly 20,000 calls a day are finished without a person

⑨2 records

Open-Weight Lineage

Oct 16, 2026Reflection — Reflection launched Beam, an open-weight model — a sparse mixture-of-experts architecture with 501 billion total and 23 billion active parameters per token, released under Apache 2.0 with a context of 1 million tokens SourceIt is the first model from a new US lab. Company-reported scores were SWEBench Verified 81.6%, GPQA Diamond 90.6%, AIME 2026 97.7% and Terminal Bench v2.1 80.1% (on BF16), and it ships in three versions, BF16, FP8 and NVFP4, with the NVFP4 checkpoint at about 280GB
Oct 15, 2026Mistral AI — Mistral AI released Timestral, a time-series forecasting model — four sizes, 5M, 28M, 123M and 209M, under a modified MIT license, producing probabilistic forecasts over 21 quantiles from a context of 32,768 time steps on CPU alone without a GPU SourceCompany-reported, the 209M model ranks No. 1 in the pretrained zero-shot categories of GIFT-Eval, TIME and fev-bench, and the 28M model beat Chronos-2, which is about four times larger. On an 8-thread server CPU the 5M model takes about 10ms and the 209M model about 140ms. A technical report was only announced

⑫2 records

Training Data & Post-Training

Oct 16, 2026NVIDIA, AI4Bharat (IIT Madras) — NVIDIA announced the coming release of a 2.8 trillion token pretraining corpus for 22 Indian languages — Nemotron-pretraining-Indic-v1, built with AI4Bharat at IIT Madras, comes together with an open generation pipeline that combines translation, quality scoring and restoration SourceCompany-reported, it covers 34 language and script combinations with 2.601 trillion tokens in native scripts plus 224.6 billion in Latin script, and includes 438.4 billion restored tokens kept only where the quality score was 75 or above. Continuing the training of Nemotron 3.5 Super VL on part of this corpus raised five Indic benchmarks by 4.98 to 9.82 percentage points, with the caveat that this is not the effect of the synthetic corpus alone
Oct 15, 2026Bingxi Hou et al. — Waypoint-OPD released — rather than taking the teacher model's change in probabilities before and after reinforcement learning as it stands, it compares them with a midpoint of the same training run and, where the direction of change before and after disagrees, damps that signal as it transfers it to the student model SourceIt starts from the point that looking only at the endpoints lets preference changes from early and late training cancel each other out and hide the parts where the teacher changed consistently. The authors reported that with a Qwen3-4B teacher and three Qwen3-family students, in all six settings compared across the two domains of math and code, it had the highest domain average among the methods compared

⑯2 records

Evaluation & Benchmarks

Oct 15, 2026Dongxu Yang et al. — FDE-Gym released — a benchmark modeled on the work of an engineer posted to a customer, with 95 delivery tasks drawn from 20 situations across 18 industries that have to be solved while asking after vague requirements and constraints nobody stated SourceThe authors reported that in the setting where the constraints have to be found by the model, even the best configuration averaged only 0.56 points and solved just 14 of the 95 tasks. Even in the 23 tasks where every necessary fact was obtained from the stakeholder, only 10 scored 0.8 or above, and they concluded that finding the unstated constraints is the bottleneck
Oct 13, 2026Xuyang Liu et al. — OmniCompass released — a benchmark that measures video understanding by splitting it into three parts, when something appears, where it is and what is happening, with a separate track that uses sight only and a track that also uses sound SourceThe authors said the vision track was built from 820 videos and 2,460 questions and the track that also uses sound from 240 videos and 720 questions, all annotated by people and checked several times, and that 50 models were evaluated. They reported that open models fell behind commercial models especially on tasks that use sound and screen together, and that answers often differed between pointing to the same object in text only and giving a picture along with it

③1 record

Tools & Agents

Oct 15, 2026Hanzhang Zhou et al. — OpenGCUA released — a study that leaves every component of a general-purpose computer-using agent open, combining screen control with code execution and moving the job of judging what to look at and how far the work has come from harness rules to the model SourceThe authors reported that on OSWorld-v2 the average partial score of the 8 models evaluated rose by 19.6 points and that steps per task fell by 53.0% on average compared with running screen control alone. They released the implementation, prompts and evaluation logs together with an early version of a harness for real devices

⑤1 record

Loop & Harness Engineering

Oct 14, 2026Cohere — Cohere published experiments that group coding agents into teams — with a minimal harness in which a lead agent hands out the work, it varied team size, model mix and whether the work ran asynchronously, and had 200 ProgramBench tasks rebuilt from scratch SourceCompany-reported, a single Sol agent at 52.1% rose to 61.8% once it worked as a team, and a team of a Sol lead with 16 Luna planners reached 61.8% for $155, about 30% of the cost of an all-Sol team that delegated every stage ($510, 63.2%). Cohere added the caveat that teams cost 3 to 7 times more than a single agent and that the score on one task swings by about 6 points from run to run

⑧1 record

Inference Infrastructure & Serving

Oct 15, 2026IBM — IBM released Granite Switch 4.1 — to handle chat, judging and speech transcription from a single endpoint, it bundles the 8B and 30B models with a judging aLoRA, a Guardian aLoRA for safety judgments and Granite-Speech-5.0-Turbo-CTC SourceBecause these are activated LoRAs, the KV cache built by the base model can be reused for the judgment that follows, and IBM said GPU memory with all three components added stays close to that of the base model alone. It gave almost no numerical comparisons, only a demonstration in which a customized 3B model judged 60 to 100 items per second

⑩1 record

Safety, Alignment & Interpretability

Oct 14, 2026Anthropic — Anthropic published its preparations for the 2026 US midterm elections — it links official voting information from Democracy Works into answers and shows the count on election day, and adds classifiers aimed at misuse split across several steps, at fake accounts and at fake news sites SourceIn a company-reported sample (Sep 21–27, 2026; about 250,000 US conversations), politics, policy and society topics made up 5.75%, elections and voting 0.51%, conversations asking who to vote for 0.01%, and election-integrity policy violations 0.002%. In its own evaluation the political balance score for Opus 5.5, Sonnet 5.5 and Haiku 5.5 was 99%, but an outside reviewer pointed to a tendency to add caveats that were not asked for and to the risk of missing an operation broken into steps that each look harmless

⑬1 record

Compute Economics & Geopolitics

Oct 15, 2026Cohere, AISLE — Cohere announced plans to acquire the security company AISLE — it agreed to buy the Czech company, founded in 2024, which builds systems that find and fix unknown vulnerabilities on their own, and set up a new Cohere Cyber organization SourceAISLE said that in its first year out of stealth it was assigned more than 500 CVEs across more than 100 codebases, and that it found all 12 vulnerabilities in OpenSSL's January security release and 18 of the 24 in curl's latest release. The amount was not disclosed, and Cohere said it will make Prague a base for AI security

⑭1 record

Physical World & Real-World Impact

Oct 15, 2026OpenAI — OpenAI released a Brazilian Portuguese course for teachers — on the free learning platform OpenAI Academy it opened an AI for Educators course for primary, middle and high school teachers, covering how to use AI for lesson design, activities, assessment and communication SourceIt put protecting teachers' educational judgment forward as a premise, and, timed to Teachers' Day, also explained how to use ChatGPT's study mode (Study) in class. No figures on the number of users or on effectiveness were given

⑮1 record

Regulation, Law & Governance

Oct 14, 2026Anthropic, Government of Singapore — Anthropic opened a Singapore office and signed an MOU with the government — with the Ministry of Digital Development and Information it agreed to work together on AI safety, security and model evaluation, research on the effects of adoption, and responsible adoption in the public and private sectors, opening its fifth office in Asia-Pacific SourceThe signing was timed to Singapore International Cyber Week, and technical exchange with local safety and assurance bodies on evaluation, testing standards and incident response is planned. Company-reported, Singapore ranks No. 2 of 121 countries in Claude.ai use per capita, and more than 1,500 GovTech engineers and technical staff use Claude Code

⑲1 record

On-Device & Local Inference

Oct 16, 2026NVIDIA — NVIDIA launched RTX Spark laptops — Windows laptops that pair Blackwell RTX graphics with its own AI platform and run models on the device with up to 128GB of unified memory, selling from $2,399 SourceASUS, Dell, HP, Lenovo, Microsoft and others are taking part in a staggered rollout. The launch put AI run on the device front and center alongside creative work and gaming, but compute performance (TOPS) was not stated in the press release

How this relates to the main moduleThis page holds only what was added that week. For the whole story, open the main module — it is one latest edition, revised every week, while weekly editions stay as they were.