NAll news

All news

Saturday41 items

  1. Rohan PaulModel releases
    AI score72

    Moonshot weakened Kimi K3.1 Pro's hacking capabilities before open release

    Before open release, Moonshot proactively weakened Kimi K3.1 Pro's cyberattack capabilities; the unweakened version had a 90.5% solve rate on its internal exploit task, while the released version was 81%. This result is still far higher than K3's 33.3%, but lower than DeepSeek V4.1-Flash's 85.7%. Moonshot said that exploit skills rise as post-training scale expands, so it removed cyber-related data from mid-training and reinforcement learning, and internal xdev-bench-hard pass@3 dropped from 90.5% to 81.0%.

    Quoted postRohan Paul · @rohanpaul_ai

    Moonshot AI’s first update, the new Kimi K3.1 Pro > Kimi K3.1 Pro scores 65.7% on Terminal-Bench 4.0, 0.7 points behind Claude Opus 5.5 and 7.5 ahead of GPT-6 Astra, at a lower average cost per task than either rival’s top runs. > Moonshot saw exploit skill climb as post-training scaled, so it removed cyber-related data from mid-training and RL; on its internal xdev-bench-hard, pass@3 fell from 90.5% to 81.0%. > The filtered model still solves 81.0% of those exploitation tasks against K3’s 33.3%, which Moonshot reads as general coding gains carrying into cyber work, while a third-party run of GPT-5.6 Sol solved 100%. > On Moonshot’s internal hallucination stress test with tools, false claims fell from 23.4% for K3 to 8.3%, still above GPT-6 Astra’s 5.5% but well below Claude Opus 5.5’s 15.0%. > The biggest jump is on Moonshot’s simulated business-operation test, where scores rose from 8.9 to 37, and K3.1 Pro also tops OfficeQA Pro v2 at 64.4 against Opus 5.5’s 62.2. > Gains are uneven: K3.1 Pro scores 38.1 on Terminal-Bench Science against GPT-6 Astra’s 68.1 and Opus 5.5’s 63.3, while leading Agents’ Last Exam at 43.2 and Humanity’s Last Exam with tools at 70.0. > K3.1 Pro and a Highspeed variant running 2-3x faster are live in Kimi’s apps and API, with full weights due by November 4 for independent reruns.

    Same story · 34 reports

    #Model releases#Safety / alignment#Open-source ecosystem

  2. Nathan LambertModel releases
    AI score70

    Kimi K3.1 released under the name 3.1 Pro, with benchmark results leading current open-weight models

    Kimi K3.1 was released under the name 3.1 Pro, and Nathan Lambert believes its benchmark scores are clearly the strongest among current open-weight models, provided the weights are released by November 4 as officially stated. The Kimi team explicitly discusses cyber risk and seeks to keep the open-weight cyber capability frontier in check, keeping it restrained relative to the closed-source frontier. K3.1's per-token price is lower than K3's, but some benchmarks in the blog show that it consumes significantly more tokens per answer, so the actual price may be similar.

    Same story · 34 reports

    #Model releases#Open-source ecosystem#Evaluations / benchmarks

  3. Nathan LambertModel releases
    AI score85

    Kimi K3.1 Pro Released, Intelligence Index Score of 55 Ranks Fourth

    Kimi K3.1 Pro scored 55 on the Artificial Analysis Intelligence Index, ranking fourth, an 11-point improvement over Kimi K3's top score of 44. The model shows notable gains in agentic capabilities, scoring 1,846 Elo on GDPval-AA, 312 Elo higher than K3, and 67% on Terminal-Bench 4.0, above K3's 13%. The model has a 1M-token context window, 2.8T total parameters and 104B activated parameters, supports text and image inputs, and is priced at $2, $10, and $0.20 per 1M input, output, and cache-hit tokens respectively, 33% lower than K3; the weights have not yet been released, and Moonshot expects to release them shortly after launch.

    Quoted postArtificial Analysis · @ArtificialAnlys

    Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index scoring 55, making large gains in agentic performance; it will be the leading open-weights model once its weights are published @Kimi_Moonshot's new reasoning model gains 11 points over Kimi K3 (max, 44). It trails only Claude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index. Once the weights for K3.1 Pro are released it will be the highest-scoring open weights model and the gap between the leading open weights and proprietary models on the Intelligence Index will narrow from 12 points to 3. Key results: ➤ Large gains in agentic performance: Kimi K3.1 Pro scores 3rd on GDPval-AA with 1,846 Elo, up 312 Elo from Kimi K3, behind only Claude Opus 5.5 (1,866 Elo) and Muse Wonder Full (1,857 Elo). Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0. K3.1 Pro scores 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 (64%). AutomationBench-AA, which tests agentic SaaS workflows, rises from 58% to 73%. ➤ Gains in other evals are more moderate: Humanity's Last Exam rises from 47% to 52%, CritPt from 23% to 25%, and SciCode from 59% to 61%. AA-Omniscience Index rises from 20 to 30, as the hallucination rate falls from 53% to 44% and accuracy rises from 48% to 51%. ➤ Cost per Task roughly equal to Kimi K3 at $1.98 per Task: K3.1 Pro uses 55K output tokens per Intelligence Index task, 14% more than Kimi K3 at 48K, while its first-party token prices are 33% lower. Additional model details: ➤ Context window: 1M tokens. ➤ Size: 2.8T total, 104B active parameters ➤ Input modalities: Text and image ➤ Weights and license: Weights currently not published, Moonshot is expected to release the weights soon after launch ➤ Pricing: $2/$10/$0.20 per 1M input/output/cache hit tokens, 33% lower than Kimi K3 on all three rates. Congratulations to the team at @Kimi_Moonshot on the release!

    #English text is being prepared.#Evaluations / benchmarks#Model releases

  4. ZhidxModel releases
    AI score75

    Ant Digital Technologies releases decision model LingDT-Decision-1.0, 27B version ranks first in overall score

    Ant Digital Technologies releases decision model LingDT-Decision-1.0, offering 4B and 27B versions, targeting high-frequency decision tasks in scenarios such as customer service, risk control, and finance, and now available for a limited-time free trial on the DTMaaS platform. The model supports native multiple-choice and sequential decision-making (DAG), and the 27B version ranks first in overall score across 23 open-source evaluation benchmarks, surpassing models such as Microsoft-Decision-1 and Jev 1.13.0. Ant Digital Technologies' tests show that in data annotation scenarios, Token consumption is reduced by about 97% and processing time improved by about 10x compared with DeepSeek-V4-flash; in intelligent customer service scenarios, Token consumption is reduced by 90% and risk identification accuracy increased from 76% to 89% compared with Qwen-3.7-flash.

    Same story · 3 reports

    #Model releases#Agents#Reasoning

  5. Unite.AIModel releases
    AI score87

    Moonshot AI releases Kimi K3.1 Pro, the first model in the K3.1 series

    Moonshot AI released Kimi K3.1 Pro on October 24, the first model in the K3.1 series, positioned as a 3T-class open-weight model supporting a 1 million token context window, and also launched Kimi K3.1 Pro Highspeed with 2 to 3 times faster speed. The company says K3.1 Pro scores 65.7% on Terminal-Bench 4.0, below Claude Opus 5.5's 66.4% and above GPT-6 Astra's 58.2%; in an internal hallucination stress test, its claim-level hallucination rate dropped from 23.4% in K3 to 8.3%. Both models are now available on Kimi.ai, mobile, the Kimi Work desktop app, Kimi Code and the Kimi API, and the full model weights will be released on November 4.

    Same story · 34 reports

    #Model releases#Open-source ecosystem#Agents

  6. Unite.AIModel releases
    AI score78

    Block releases full-duplex speech model Dessa, trained from scratch using only human conversation data

    Block's internal AI research lab released the full-duplex speech-to-speech model Dessa on October 23, stating that it was trained entirely from scratch on human conversational speech, can listen and speak simultaneously, and natively perceives emotion, sarcasm, and laughter. The model's inference speed is faster than real time, and it can run locally on a MacBook Pro M4 Max or higher configuration; the team completed training with a single NVIDIA GB200 NVL72. Block has published demo recordings and a public waitlist, and has opened early access to a small number of AI safety researchers.

    Same story · 10 reports

    #Model releases#Speech#Open-source ecosystem

  7. The Verge AIModel releases
    AI score58

    Block releases pure speech dialogue model Dessa

    Jack Dorsey's Block releases AI model Dessa, claiming it is trained entirely from scratch on human conversations and is a true speech-to-speech model. The official release showcases several dialogue samples, including a comparison with OpenAI's GPT-Live-1 speech model, in which Dessa appears more natural. Dessa can run locally on a MacBook Pro M4 Max or higher configuration, and the team says it will open early access to a small group of AI safety researchers.

    Same story · 10 reports

    #Model releases#Speech#Open-source ecosystem

  8. Full article from Artificial AnalysisFirst-partyModel releases
    AI score78

    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.

    Same story · 34 reports

    #Model releases#Reasoning#Agents

  9. MarkTechPostModel releases
    AI score82

    Moonshot AI releases Kimi K3.1 Pro: 2.8T total parameters, 104B activated, 1M context

    Moonshot AI releases Kimi K3.1 Pro, the first model in the K3.1 series, using a MoE architecture with 2.8T total parameters, 104B activated, and support for a 1M token context window. The model is now available on Kimi.ai, Kimi Work, Kimi Code, and the OpenAI-compatible Kimi API, with API pricing at $2 per million tokens for input and $10 for output, lower than K3's $3 and $15. Moonshot self-reports a Terminal-Bench 4.0 score of 65.7%, and SWE-Marathon 51.7 leads all models in its chart, but Terminal-Bench Science 0.1 is only 38.1, behind GPT-6 Astra's 68.1; open weights are expected to be released in November 2026.

    Same story · 34 reports

    #Model releases#Open-source ecosystem#Agents

  10. WallstreetcnModel releases
    AI score85

    Moonshot AI releases Kimi K3.1 Pro, 2.8 trillion parameters, coding agent benchmark surpasses GPT-6 Astra

    Moonshot AI launches flagship model Kimi K3.1 Pro, with 2.8 trillion total parameters, adopting MoE architecture, 104 billion parameters activated each time, supporting a 1 million Token context window, and plans to open full model weights before November 4. The model scored 65.7% on the Terminal-Bench 4.0 coding agent test, higher than GPT-6 Astra's 58.2%, slightly lower than Claude Opus 5.5's 66.4%; on the BenchCAD benchmark it scored 92.7% with tool use. The standard version API output price is 68 yuan per million Tokens; the high-speed version is about 2 to 3 times faster than the standard version and priced at twice the price, and has been integrated into the Kimi official website, App, Kimi Work, Kimi Code and API.

    Same story · 34 reports

    #Model releases#Agents#English text is being prepared.

  11. Moonshot AI · Kimi BlogFirst-partyModel releases
    AI score87

    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%.

    Benchmark results for Kimi K3.1 Pro and peer models across 14 coding, science, and knowledge-work evaluations

    Same story · 34 reports

    #Model releases#Open-source ecosystem#Agents

  12. Google BlogFirst-partyModel releases
    AI score80

    Google releases Gemini 4 Argon, showcasing 5 community creative projects

    Google releases new frontier model Gemini 4 Argon, stating it performs outstandingly on complex tasks such as coding, mathematics, and 3D world building, and surpasses other frontier models on benchmarks across multiple domains. The official showcase features 5 community projects, including a 3D Earth that maps wildfires and seismic activity in real time, a turbofan engine simulation with real-time data, a 3D carriage generated from a single image, and a multiplayer capture-the-flag browser game. The model has been widely available since last week.

    Collage showing five builds with Gemini 4 Argon

    Same story · 2 reports

    #Product updates#Google#Multimodal

  13. Simon WillisonModel releases
    AI score81

    Moonshot AI's Kimi K3.1 Pro launches on API, weights to be opened on November 4

    Kimi K3.1 Pro launches on the Kimi API today; weights are not yet open, and Moonshot AI promises to release the full model weights before November 4, 2026. Pricing is $2.00 per million input tokens and $10 per million output tokens, the same as GPT-6.1 Sol and half that of Claude Opus 5.5; a Highspeed version that is 2-3 times faster at double the price is also offered. Artificial Analysis gives Kimi a score of 55, higher than Claude Fable 5.1, GPT-6 Astra and Gemini 4 Argon at 53, and second only to Claude Opus 5.5 at 58. Simon Willison tested SVG generation via the Moonshot API at three reasoning effort levels: low, medium and max.

    Same story · 34 reports

    #Model releases#DeepSeek#Reasoning

Friday28 items

  2. LeiphoneModel releases
    AI score63

    Kimi K3.1 leak: 3D generation cost is about 1/14 of GPT-6 Astra

    Leiphone summarized the leaked information about Kimi K3.1: the text "Kimi K3.1" appeared in Kimi's official teaser; developers had previously found the kimi-k3-1 identifier in the Moonshot AI API configuration and the Kimi Code model list, and the anonymous arena model Kibitz is speculated to be this version. The article states that it performs outstandingly on 3D and front-end code tasks such as Three.js, that the token consumption for generating 3D code of equivalent complexity is far lower than that of top foreign models, with a cost of about 1/14 of GPT-6 Astra, and that a new "Medium" thinking tier has been added, expanding the thinking intensity from three levels to four. The article also mentions that it has obvious shortcomings in generating biological motion trajectories such as a running giraffe.

    #English text is being prepared.#Multimodal#English text is being prepared.

  4. Testing CatalogModel releases
    AI score85

    MiniMax publicly releases M3.1 Flash weights on Hugging Face

    MiniMax has publicly released the MiniMax M3.1 Flash weights on Hugging Face; the model has 428B total parameters and 23B activated per token. The company claims generational improvements in agent tasks, coding, and cybersecurity, with scores of 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, and 82.2 on CyberGym. This model is the stealth model Space Bunny on OpenRouter, released on October 19, and is now available on the MiniMax platform and API.

    Quoted postMiniMax (official) · @MiniMax_AI

    MiniMax M3.1 Flash weights are now public. 428B parameters total, 23B activated per token. Agentic performance: A generational leap in agentic tasks, coding, and cybersecurity — 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, and 82.2 on CyberGym. Efficient by design: MiniMax Sparse Attention and DSpark speculative decoding improve inference efficiency. Hugging Face: https://t.co/JDYA1zZh54 ModelScope: https://t.co/DQcmOjTwhO API: https://t.co/fHRdSV7BwZ M Plan: https://t.co/LYnYZcsesu MCode for desktop: https://t.co/SGmLTI3fFU

    Same story · 5 reports

    #Model releases#Open-source ecosystem#Hugging Face

  5. MiniMaxFirst-partyModel releases
    AI score65

    MiniMax Opens M3.1 Flash Model Weights

    MiniMax has announced that the model weights for M3.1 Flash are now open. Together AI, a day-one launch partner, has deployed the model in its AI Native Cloud production environment, so users do not need to manage GPUs themselves. According to Together AI, M3.1 Flash targets agentic tasks, coding, computer use, and cybersecurity scenarios, and supports a 1M context window and vision capabilities.

    Quoted postTogether AI · @togethercompute

    minimax m3.1 flash is now live on together ai we’re excited to be a day-0 launch partner for the model, aka space bunny m3.1 flash is built for agentic tasks, coding, computer use + cybersecurity, with a 1m context window and vision support try it now https://t.co/sdIawJuwm4 https://t.co/SSgBvYEiGY

    Same story · 5 reports

    #Model releases#Open-source ecosystem#English text is being prepared.

  6. TMTPost · AGIModel releases
    AI score71

    Mirendil releases a hundred-billion-parameter coding foundation model Mirendil-Base, trained by 1 researcher and an Agent

    Mirendil releases a coding-specific foundation model Mirendil-Base, with a parameter range of 100 billion to 500 billion, completed by 1 researcher and an Agent in three and a half months, at a cost of less than 10 million USD, with over 99% of the code written by the Agent. The official statement says training used entirely public data such as GitHub open-source code, public web pages, and public datasets, with no purchased data, no distillation, and no synthetic data. In the cross-entropy test on new code repositories, Kimi K3 ranks first with 0.193, Mirendil-Base ranks second with 0.222, followed by GLM-5.3, DeepSeek-V4, and Nemotron; the article also points out that foundation models are usually better than post-trained models in this test, so the actual gap between Mirendil-Base and Kimi K3 is larger than the scores show.

    Same story · 2 reports

    #Model releases#Agents#English text is being prepared.

  7. The PaperModel releases
    AI score73

    MiniMax open-sources the M3.1 Flash model, reducing usage cost by 70%

    MiniMax announced on October 23 the open-sourcing of the M3.1 Flash model, which processed over 100 trillion tokens within 72 hours of launch. Compared with the previous generation M3, M3.1 Flash has comprehensively improved capabilities, with usage cost reduced by 70% and online inference speed increased by nearly 3 times. MiniMax stated it will work with chip manufacturers such as Huawei Ascend, NVIDIA, and AMD, as well as inference frameworks and communities such as vLLM, SGLang, and Hugging Face, to advance day-one open-source integration.

    Same story · 5 reports

    #Model releases#Open-source ecosystem#Agents