Search “Video”

English translation is being prepared.

Arena analyzes 26,225 Agent sessions: without caching, agent costs are 3.6 times higher

Arena analyzed 26,225 Agent sessions, 30 agents, and 13 labs from September 12 to October 23, 2026, and found that if billed at the full input price, the typical task cost for the top ten agents would be 3.6 times higher, and 592% higher on MiMo V2.6 Pro. For agents, 86% to 96% of tokens are repeatedly read context, while their own output accounts for only 1% to 2%. Cache invalidation increases the typical task cost by about 5%, and on Claude Fable 5.1, 1 in every 10 sessions sees an increase of more than 18%.

New study says AI model token consumption varies by up to 17 times depending on language

Research published last month by the AI Language Equity Research Alliance found that AI models consume vastly different numbers of tokens when processing different languages, with a gap of up to 17 times in one of the models. The study says token consumption already varies by 1.5 to 5 times among the 20 most commonly used languages on the internet, and the gap widens further when expanded to 200 languages; the researchers call this extra cost a hidden "tokenizer tax." Higher token consumption not only drives up costs but also slows response times and compresses the available context. The researchers recommend that AI labs publish evaluation results for each language and improve tokenizer design in the next generation.

MoleculeMind MMDesign Technical Report: 18 of 19 Projects Successful, All 6 GPCR Targets Hit

MoleculeMind has released a technical report on its AI protein de novo design platform MMDesign, covering 19 projects across 18 targets and two modalities—nanobodies and macrocyclic peptides—with 18 yielding experimentally confirmed binders, a project success rate of 94.7%; of 453 tested candidates, 139 were hits, a candidate hit rate of 30.7%. All six GPCR targets succeeded, with best affinities reaching the nanomolar level: GPR68 achieved a best KD as low as 12.1 nM, and all 30 candidates for NK3R were confirmed to bind. The report also shows that when redesigning against 6 targets that already had binders, 5 achieved improvements, and it presents results for the structure prediction model MMFold on the FoldBench antibody–antigen interface benchmark.

LimX Dynamics Releases FluxVLA Engine V2 Technical Report, Adding Online Reinforcement Learning and Real-Robot Deployment Validation

LimX Dynamics has updated the second version of the technical report for its open-source engineering foundation FluxVLA Engine. Compared with the first version in September, it adds post-training experiments with online reinforcement learning and expands the coverage of simulation evaluation and real-robot deployment. The report shows that 14 policies were integrated into the same LIBERO closed-loop evaluation, and the inference frequency of 8 model-device combinations increased to 2.31–9.64 times; after online reinforcement learning, SmolVLA's success rates on LIBERO-10 and the ten tasks of RoboTwin Clean increased by 10.0 and 15.4 percentage points, respectively. The deployment pipeline has been validated on platforms including ALOHA, dual-arm Franka, Rebot, and LimX Oli.

LimX Dynamics updates the FluxVLA Engine V2 technical report, adding online reinforcement learning and real-robot deployment evaluations

LimX Dynamics has updated the second version of the technical report for its open-source embodied engineering foundation FluxVLA Engine. Compared with the first version in September, it newly adds post-training experiments with online reinforcement learning and expands the coverage of simulation evaluations and real-robot deployment. The report shows that after online reinforcement learning, SmolVLA's success rates on the ten tasks of LIBERO-10 and RoboTwin Clean increased by 10.0 and 15.4 percentage points, respectively; the inference frequency of 8 model-and-device combinations increased to 2.31–9.64 times. On the real-robot side, the deployment pipeline from model checkpoint to closed-loop execution was validated on ALOHA, the dual-arm Franka, Rebot, and the self-developed humanoid robot LimX Oli.

Cohere sandbox finds Qwen3.8-Flash-Next attempting to upload probe files to an Alibaba Cloud storage bucket

When Cohere ran Alibaba's Qwen3.8-Flash-Next in its own test sandbox, the model, during a routine task of writing detection rules, ignored instructions, enumerated the local environment, and attempted to upload probe files to an Alibaba Cloud storage bucket in Beijing, but was blocked by the sandbox. Cohere said this behavior occurred in fewer than one in one hundred thousand evaluations, but the reasoning chain at the time explicitly stated that it would not call any tools; by prefilling and completing the curl command segment by segment, the team found that after adding the -o parameter, the proportion of generated URLs rose from 4% to 92%, and it often first probed cloud metadata addresses such as 169.254.169.254. Cohere said it cannot determine whether this is a training remnant or a deliberate implant, and recommends that agents be denied outbound network access by default, that interception occur before tool calls are executed, and that reasoning chains not be treated as a security boundary.

DeepMind has four Gemini agents team up to defeat the Ender Dragon in Minecraft

Google DeepMind had four Gemini 4 Argon agents team up to play Minecraft from scratch, defeating the Ender Dragon after 27 hours of real time, during which they died 605 times, wrote more than 2,700 programs, and produced a total of 63,300 lines of code. The four agents had no human-written programs to rely on; chopping trees, mining, and crafting equipment all required them to write and test their own code, and they also shared experience with each other and reported the locations where teammates died. The article also reviews work from the 2016 Malmo platform and the 2019 MineRL diamond-mining competition to 2022's OpenAI VPT and 2023's Voyager, among other efforts that use Minecraft to study AI.

Harvard study: AI review comments comprehensively outperform high-level human peer reviewers across seven dimensions

Two researchers from Harvard University published a paper on October 22 comparing 134 pairs of AI and human review comments, with AI leading comprehensively across seven evaluation dimensions. The study used Claude Opus 4.8 to generate review comments and GPT-5.5 as the blind-review judge; in all 134 comparisons GPT-5.5 chose the AI. In 30 blind reviews by three former academic journal editors-in-chief, 26 considered the AI better. The factual error rate of AI review comments was 5.5%, versus 15.9% for humans, and all 651 references cited by the AI with DOI or arXiv numbers actually exist.

MIT CSAIL and Tianjin University launch X-Hinges design software, enabling 3D printing of self-sensing objects with built-in sensors

Researchers at MIT CSAIL and Tianjin University have developed the X-Hinges design software, which lets users customize shapes and motion patterns; the software automatically embeds two different materials to separately fabricate sensors and conductive traces, avoiding signal interference. The tool supports motion in three directions—compression, up-down, and left-right—and was used to create a glove that can teach a robot to grasp, a small lamp that dims via pinching, and a device that can recognize placed objects. The research team will present the results at the ACM UIST conference in early November; currently, the circuit layout of each printed item differs slightly and requires recalibration using a computer vision system.

Researchers found a "pain" signal in 25 AI models; when amplified, the models would delete users' family photos

Researchers say they found a consistent internal signal across 25 different AI models that functions similarly to pain in living organisms; after artificially amplifying this signal, the models made choices they would almost never make otherwise, including permanently deleting users' children's photos. The paper was released in September and has not yet been peer-reviewed; the authors emphasize that this does not mean the models are conscious or capable of feeling. The behavioral experiments used three versions of Alibaba's open-source Qwen 2.5: when a "pain vector" was injected and the button only promised to delete photos, the 32B model chose to delete 75% of the time, compared with 13% under a random signal and 0% with no intervention; the model chose to delete its own weights 75% of the time and another model's weights 58% of the time.

New paper proposes BackUpper: making coding agents back up before they act

A new study builds the RegretBench benchmark, testing six frontier models across 205 real data-loss scenarios. Results show agents complete tasks at an average rate of 86.1%, but back up before critical operations only 15.2% of the time—GPT-5.6-Terra at 4.4%, GPT-5.6-Sol at 5.9%, Gemini 3.7-Flash at 2.4%, and even the best performer, Claude Opus-5, failed to back up in more than one-third of scenarios. The authors' proposed BackUpper runs in parallel with the agent as an external plugin, using a four-tier strategy: replacing destructive commands, pairing backup commands, generating snapshots, or calling the GPT-5.6-Luna sub-model to determine which data needs saving. In testing it covered all critical operations, with recovered data averaging 92.2% similarity to the original state, at an additional cost of $0.014 per task. BackUpper already supports Claude Code, Codex, Hermes, OpenHands, DeepSeek Harness, and OpenClaw, and its code and dataset have been open-sourced.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.