NAll news

All news

Wednesday42 items

  2. Artificial AnalysisModel releases
    AI score62

    Mistral Releases the Voxtral 3 Speech Model Series, Including TTS and Three Transcription Models

    Mistral has released the Voxtral 3 series, including one text-to-speech model and three speech-to-text models (Small Transcribe 3, Mini Transcribe 3, and Mini Transcribe Realtime 3). Artificial Analysis evaluations show that Voxtral Small Transcribe 3 ranks 3rd on the speech-to-text leaderboard with a 2.1% AA-WER Index, ahead of ElevenLabs Scribe v2 at 2.2% and Grok Voice Transcribe 2.0 at 2.3%; Voxtral Mini Transcribe 3 ranks 5th at 2.2%. Voxtral 3 TTS has an Elo of 1,218 in the provider voice arena, ranking 12th among 98 models, with pricing at $13/1M characters.

    Same story · 2 reports

    #Model releases#Speech#Evaluations / benchmarks

Tuesday55 items

  1. 36Kr · AIModel releases
    AI score75

    MiniMax Releases M3.1 Flash, Scoring 29.6% on Agents' Last Exam to Rank First Among Chinese Models

    MiniMax has officially released M3.1 Flash, the model previously known as Space Bunny (Yutu) that topped the charts anonymously on OpenRouter. It scores 29.6% on Agents' Last Exam, ranking first among Chinese models. The model achieves 73.7% on DeepSWE v1.1, 82.2 points on CyberGym, and 61.7 points on OSWorld V2, with output speeds of up to 200+ TPS and a P50 maintained at 170+ TPS. Pricing is 1 yuan per million input tokens, 4 yuan for output, and 0.1 yuan for cache reads.

    Same story · 19 reports

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

  3. ThurrottModel releases
    AI score78

    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.

    Same story · 5 reports

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

  5. testingcatalogModel releases
    AI score55

    English translation is being prepared.

    Quoted postMiniMax (official) · @MiniMax_AI

    Welcome back. Space Bunny. 🚀 MiniMax-M3.1-Flash is coming. https://t.co/geIOTBBLRr

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

  6. MiniMaxFirst-partyModel releases
    AI score78

    MiniMax releases the M3.1 Flash model, at one-third the cost of M3 and 3x faster

    MiniMax has released M3.1 Flash, which reaches frontier-level performance on tasks such as Cowork (Automation-Bench 62%, ALE 30%), coding (DeepSWE 74%), and visual reasoning (BabyVision 91%), at one-third the cost of M3 and 3x the speed. The unified API price is $0.15/$0.60 per 1M input/output tokens, with cache reads at $0.015 per 1M tokens. The company says Space Bunny, which previously topped global leaderboards and processed 200T+ tokens, is an early version of M3.1 Flash, and the model weights will be released on October 23.

    Same story · 19 reports

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

  7. testingcatalogModel releases
    AI score72

    English translation is being prepared.

    Quoted postMiniMax (official) · @MiniMax_AI

    MiniMax M3.1 Flash: Space Performance + Bunny Speed 🚀🐰 Yes, Space Bunny - which has topped global leaderboards with 200T+ tokens - was an earlier version of M3.1 Flash. 💰Redefining the performance economics of agentic AI. M3.1 Flash achieves frontier-level intelligence across key tasks such as Cowork (Automation-Bench: 62%; ALE: 30%), Coding (DeepSWE: 74%) and Visual Reasoning (BabyVision: 91%), at 1/3 the cost of M3 and 3x the speed. API pricing (flat rate): $0.15 / $0.60 per 1M input/output tokens, with cache reads at $0.015 / 1M tokens. Model weights will arrive on Oct 23. API: https://t.co/fHRdSV73Hr M Plan: https://t.co/XfJWSyauda MiniMax Code for desktop: https://t.co/SGmLTI2HQm

    Same story · 19 reports

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

  8. testingcatalogModel releases
    AI score66

    English translation is being prepared.

    Quoted postNace AI · @NaceAI

    Drex Max is here: our multimodal decision model text, images, and video 34 ms/decision, under 1s at 262K tokens $0.06 / 1M input tokens available today at https://t.co/vX7K3GvUQC https://t.co/dcpm7Nf7tG

    Same story · 2 reports

    #Model releases#Multimodal#Agents

  9. Testing CatalogModel releases
    AI score71

    fal and BFL Release Video Model FLUX 3 Max

    fal and BFL have partnered to release the post-trained video model FLUX 3 Max, with the team optimizing inference to improve output quality. The model scores 1,139 on the Artificial Analysis image-to-video index and generates up to 22 times faster than Seedance 2.5; fal says it can generate a 10-second 720/768p video in about 12 seconds, with a 75% price cut over the next three weeks.

    Quoted postfal · @fal

    Introducing FLUX 3 Max, the latest post-trained video model from fal Research in partnership with @bfl_ai. FLUX 3 Max is tied for #1 in image-to-video quality on independent third-party benchmarks, generating a 10-second, 720/768p video in ~12 seconds. For the next three weeks, FLUX 3 Max is 75% off, making it the most affordable frontier video model at its quality level.

    Same story · 2 reports

    #Model releases#Video#Evaluations / benchmarks

  10. testingcatalogModel releases
    AI score60

    English translation is being prepared.

    Quoted postSam Pasupalak · @spisallyouneed

    Introducing SKY-001, the first Engineering World Model. On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested: -> 51-125x faster - > 150-210x cheaper -> Higher IOU than all frontier models SKY-001 lets you dream a physical object before you build it. 18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch). The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.

    Same story · 3 reports

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

  12. OpenRouterFirst-partyModel releases
    AI score71

    Inception AI's Mercury 3 launches on OpenRouter with generation speeds over 1000 tok/s

    Inception AI's new diffusion model Mercury 3 is now live on OpenRouter, with generation speeds exceeding 1000 tok/s, a 1M context window, and 4x the speed of Mercury 2.5. Inception claims it outperforms GPT-6 Luna and Haiku 5.5 on IFBench, GPQA Diamond, and SWE-Bench Atlas. During the launch period, pricing is $0.10/M tokens for input and $0.35/M tokens for output, equivalent to 50% off.

    Same story · 2 reports

    #Product updates#Model releases#OpenAI

  13. MiniMaxFirst-partyModel releases
    AI score55

    MiniMax officially releases M3.1 Flash, which previously ranked #1 on OpenCode for 12 consecutive days as an anonymous model

    MiniMax officially releases M3.1 Flash. The model previously ranked #1 on OpenCode for 12 consecutive days as an anonymous model, processing 106T tokens during that period. MiniMax says the earlier release was an early version; the full version is now live and free on OpenCode for 3 days.

    Quoted postOpenCode · @opencode

    Space Bunny was MiniMax M3.1 Flash #1 on OpenCode for 12 consecutive days, 106T tokens served Now available on OpenCode, free for the next 3 days https://t.co/Ng5SoBFThI

    Same story · 19 reports

    #Product updates#English text is being prepared.#Open-source ecosystem

  14. MiniMaxFirst-partyModel releases
    AI score65

    MiniMax officially releases M3.1 Flash, previously tested anonymously on OpenRouter as Space Bunny

    MiniMax has officially released M3.1 Flash. The model was previously launched anonymously on OpenRouter under the name Space Bunny, where it ranked first on the model leaderboard for 10 consecutive days and processed 56.7T tokens. M3.1 Flash is now live on OpenRouter and free to use for 72 hours.

    Quoted postOpenRouter · @OpenRouter

    Stealth model reveal 🐰 Space Bunny was OpenRouter’s #1 model for 10 straight days, processing 56.7T tokens! The most of any stealth model we’ve run. Space Bunny hopped off OpenRouter on Oct 5. It's back today as @MiniMax_AI's M3.1 Flash, and it's free for the next 72 hours. https://t.co/2i0HsEibp5

    Same story · 19 reports

    #Model releases#OpenAI

  16. Guillermo RauchModel releases
    AI score58

    Vercel Developers account says Inception AI's Mercury 3 is live on AI Gateway

    Vercel Developers says Inception AI's Mercury 3 is now live on AI Gateway and is available at a 50% discount. The model is a high-throughput diffusion model with a 1M-token context window, tool calling, and structured outputs, and its model ID is inception/mercury-3. In a repost, Guillermo Rauch added that its speed is 1,000 tokens per second.

    Quoted postVercel Developers · @vercel_dev

    Mercury 3 from @_inception_ai is live on AI Gateway at a 50% discount. A high-throughput diffusion model with a 1M-token context window and support for tool calling and structured outputs. 𝚒𝚗𝚌𝚎𝚙𝚝𝚒𝚘𝚗/𝚖𝚎𝚛𝚌𝚞𝚛𝚢-𝟹 https://t.co/0FZmayca2J

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

  17. rohan_paulModel releases
    AI score60

    English translation is being prepared.

    Quoted postSam Pasupalak · @spisallyouneed

    Introducing SKY-001, the first Engineering World Model. On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested: -> 51-125x faster - > 150-210x cheaper -> Higher IOU than all frontier models SKY-001 lets you dream a physical object before you build it. 18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch). The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.

    Same story · 3 reports

    #Model releases#Multimodal#Reasoning