NAll news

All news

Friday1455 items

  2. Yuchen JinOpinion
    AI score65

    Yuchen Jin: Artificial Analysis data shows Gemini 4 Carbon is smarter than…

    Citing Artificial Analysis data, Yuchen Jin says Gemini 4 Carbon is smarter than Opus 5.5 and costs less per task. He raises one caveat: the evaluation ran Gemini 4 Carbon using the Antigravity CLI, and he hasn't used Antigravity for months and isn't in a hurry to install it; he plans to try it in Codex and Omnigent instead to see how it actually performs.

    #English text is being prepared.#Evaluations / benchmarks#Google

  3. Yuchen Jin
    AI score22

    Yuchen Jin Calls Out Snowflake: DeepSeek Model Already Three Months Behind

    Yuchen Jin says Ali and Snowflake CEO Sridhar staged a "Kendrick vs. Drake"-style showdown on X, and called out Snowflake: welcoming them to integrate Unity Gateway and their own AI products, adding that there are "no hard feelings," but noting that Snowflake's latest DeepSeek model is already a 3-month-old version, while they released their OSS model on Day 0.

    #English text is being prepared.#Open-source ecosystem#DeepSeek

  6. Saito Hardcore HackerOpinion
    AI score81

    SemiAnalysis Analyzes Anthropic S-1: 90% Incremental Gross Margin and Half-Trillion Compute Commitment

    The SemiAnalysis podcast analyzes Anthropic's S-1, pointing out its 90% incremental gross margin, 40% incremental EBIT, 85% inference gross margin, and 41% compute gross margin, and that it has been profitable for two consecutive quarters on a non-GAAP basis. On customer net revenue retention, customers who spent $1 last year now spend $9, with the Q1 figure at 500%; in addition, 250 customers pay over $10 million annually and 1,900 pay over $1 million. The episode also breaks down the $500 billion to $600 billion compute commitment, which at $100 billion per gigawatt translates to only about 5 gigawatts, and discusses the discrepancy in reporting that the top five customers account for 25% of revenue and that 48% goes through cloud vendor marketplaces.

    #Industry news#OpenAI#Anthropic

  8. Ethan Mollick
    AI score40

    Historian builds neuron atlas for Talkie-1930

    Ethan Mollick noted that there have recently been many interesting experiments involving activating individual neurons in open AI models and locking them into specific personas. Historian @ResObscura tried the same method on a model trained only on pre-1930 data, with interesting results; he himself said that, inspired by recent persona research from @Sauers_, @jfonsecarivera and others, he built a partial neuron atlas for Talkie-1930, the historical language model created by @status_effects and @AlecRad.

    Quoted postBenjamin Breen · @ResObscura

    Inspired by @Sauers_ and @jfonsecarivera and other fun recent work with personas in AI models, I built a partial neuron atlas for Talkie-1930, the historical language model created by @status_effects and @AlecRad. Here's the link: https://t.co/ROz61DW4RW https://t.co/8RvqEsx5K0

    #English text is being prepared.#Reasoning

  10. Mistral AI · X
    AI score36

    Voxtral Mini Transcribe 3 ranks second on the Voice Arena speaker diarization leaderboard

    Mistral AI reposted the Voice Arena evaluation: Voxtral Mini Transcribe 3 ranks 2nd by DER (40.4%) among 9 models on the Diarization + ASR leaderboard, outperforming the larger Voxtral Small Transcribe 3 in the same family (41.1%, 4th).

    Quoted postVoice Arena · @voicearena_ai

    🚀 Mistral AI’s Voxtral Mini Transcribe 3 ranks #2 out of 9 models by DER on Voice Arena’s Diarization + ASR leaderboard, outperforming its larger counterpart. Following the launch of Mistral’s new Voxtral 3 family, we evaluated its Mini and Small transcription models on Voice Arena’s Diarization Bench. 🥈 Voxtral Mini Transcribe 3: 40.4% DER (#2) Voxtral Small Transcribe 3: 41.1% DER (#4) Both models were tested on 139 held-out conversations under identical conditions, using human-verified references, a strict 0 ms collar, and no predefined speaker counts or audio cleanup. Mini leads on in-person conversations (40.6% vs. 41.5%), while Small performs slightly better on online calls (39.1% vs. 39.5%). Overlapping speech remains the biggest challenge. Missed speech accounts for 82% of Mini's total DER. Both models correctly identify the exact number of speakers in 83.5% of sessions. Congratulations to @MistralAI on the release ! 👏 View complete track & condition breakdown: https://t.co/sLGwgH3EWt

    #Evaluations / benchmarks#Speech

  11. Mistral AI · XProduct release
    AI score51

    Mistral Launches Public Preview of Responses API for Agent Building

    Mistral has released a public preview of the Responses API for building agents on Mistral. Each response returns the model's answer, reasoning process, and tool calls as independent items, making it easier for applications to process them step by step than with Chat Completions. The API is available for all chat models hosted by Mistral, including Large 4.

    #Product updates#Agents#English text is being prepared.

  14. Sundar PichaiModel releases
    AI score82

    Google releases Gemini 4 Carbon frontier model, claims first place on Intelligence Index

    Google releases its new-generation frontier model Gemini 4 Carbon, just two weeks after the release of Argon. The model ranks first with 59 points on the updated ArtificialAnlys Intelligence Index v5.0, and achieved scores of SEC-bench Pro 88.2%, ExploitGym 54.8%, and CWE-bench 71.0% in cybersecurity evaluations. Compared with Argon, Carbon shows significant improvement on long-horizon tasks, and frontier safety safeguards have also been strengthened; starting this week, it will first be opened to select trusted cybersecurity defenders, then expand the range of partners through the Fairwind Program, and gradually open up more broadly.

    Same story · 34 reports

    #Model releases#Google#Evaluations / benchmarks

  16. Nathan Lambert
    AI score36

    Nathan Lambert on benchmarks: open-source models are not declining, the AA index previously overestimated them

    Nathan Lambert believes the decline of open-source models has been exaggerated, but the previous AA index did indeed overestimate them somewhat. He argues that the industry needs a more granular intelligence index, because LLMs are used for vastly different tasks; this is his meta-evaluation of the benchmarking industry.

    Same story · 3 reports

    #English text is being prepared.#Evaluations / benchmarks

  17. Nathan Lambert
    AI score23

    Nathan Lambert: Private benchmarks and open-source benchmarks each have their trade-offs

    Nathan Lambert pointed out that private benchmarks often show open-source models lagging further behind, but the problem with private benchmarks is that there is no community to help find and fix the inevitable bugs. He said he holds a 50/50 view on whether private benchmarks or open-source benchmarks are the better approach.

    #English text is being prepared.#Evaluations / benchmarks#Open-source ecosystem

  18. Nathan LambertOpinion
    AI score76

    Nathan Lambert on not knowing how to try Gemini in the agent era

    Nathan Lambert said he does not know how to try Gemini in the agent era and believes he needs to make up for this. He mentioned that Gemini seems to have been absent since December 2025, but may be making a comeback. The tweet quoted Sundar Pichai's post announcing Gemini 4 Carbon, which said the model ranks first with 59 points on the Artificial Analysis Intelligence Index v5.0, and scored 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench.

    Quoted postSundar Pichai · @sundarpichai

    I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model. Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching. We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards. As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!

    #English text is being prepared.#Agents#Google