Skip to main content
ShippedNo vapor

The AI that actually shipped

Change is the items in the last 4 weeks against the 4 weeks before. There is no percent when those earlier 4 weeks hold fewer than 5 items or the topic has under 8 weeks of history; under 4 weeks it is new.

Latest

Newest first. Scores are the editor’s, and the editor’s score is the headline figure. Upvotes and ratings from readers are counted apart, on each item. How ranking works →

  1. 1
    REAX 1.1: a 27B decision model fine-tuned from pplx-decider
    huggingface.co§ Decisions in the wildREAX (decisionmodels)21 Oct
    makerREAXtasktyped decisions
    Hype gap ▼ −0.8 under
    Shipped
    Real3.5
    Useful3.0
    Notable2.5
    Novel2.5
    AI first pass
  2. 2
    Decision4Rec releases Kev models that rerank recommender candidates
    huggingface.co§ Decisions in the wildMarco Polignano (m-polignano)19 Oct
    makerMarco Polignanotaskrecommender candidate ranking
    Hype gap ▬ 0.0 level
    Shipped
    Real3.5
    Useful3.0
    Notable2.5
    Novel4.0
    AI first pass
  3. 3
    AccuKnox Inc. releases Guard 1.1, an Apache-2.0 guardrail for agent inputs and prompt injection
    huggingface.co§ Decisions in the wildAccuKnox Inc. (Accuknoxtechnologies)16 Oct
    makerAccuKnox Inc.taskprompt injection defencetaskcontent moderation
    Hype gap ▼ −1.0 under
    Shipped
    Real4.0
    Useful3.5
    Notable2.5
    Novel3.0
    AI first pass
  4. 4
    Maincode releases Matilda Decision v1, a 27B multimodal decision model
    huggingface.co§ Decisions in the wildMaincode (Maincode)15 Oct
    makerMaincodetasktyped decisionstaskmultimodal classification
    Hype gap ▼ −0.5 under
    Shipped
    Real3.5
    Useful3.5
    Notable3.0
    Novel3.0
    AI first pass
  5. 5
    decisionmodels releases s1-llm-auto-router, a 183 MB CPU router under MIT
    huggingface.co§ Decisions in the wilddecisionmodels (decisionmodels)15 Oct
    makerdecisionmodelstaskLLM request routing
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel3.0
    AI first pass
  6. 6
    Liquid AI releases LFM2.5-24B-A2B, a 24B on-device reasoning MoE with 2B active
    huggingface.co§ Open weightsLiquid AI (LiquidAI)12 Oct
    makerLiquid AItaskon-device agentstasktool use
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.5
    AI first pass
  7. 7
    StartLux releases Decision v1.1, from 0.8B to 27B, with non-commercial weights
    huggingface.co§ Decisions in the wildStartLux (startlux-models)11 Oct
    makerStartLuxtasktyped decisions
    Hype gap ▼ −0.5 under
    Shipped
    Real3.5
    Useful3.5
    Notable3.0
    Novel3.0
    AI first pass
  8. 8
    AI Singapore releases Gemma-SEA-LION-Guard v2, MIT-licensed safety models for Southeast Asian languages
    huggingface.co§ Decisions in the wildAI Singapore (aisingapore)11 Oct
    makerAI Singaporetaskcontent moderationtaskaction governance
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.5
    AI first pass
  9. 9
    Microsoft releases Decision-1, a 9B model that scores fixed answer options
    commandline.microsoft.com§ Decisions in the wildMicrosoft9 Oct
    makerMicrosofttaskdecision scoringtaskrouting and classification
    Hype gap ▼ −0.2 under
    Shipped
    Real3.5
    Useful4.0
    Notable4.0
    Novel3.0
    AI first pass
  10. 10
    Amazon publishes Strands Decider 2B, a small decision model for agent workflows
    huggingface.co§ Decisions in the wildAmazon (amazon)9 Oct
    makerAmazontaskagent routingtaskguardrails
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  11. 11
    Reflection releases Beam, a 501B open-weights MoE under Apache 2.0
    huggingface.co§ Open weightsReflection (reflection)7 Oct
    makerReflectiontasksoftware engineeringtaskagentic tool use
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.5
    Notable4.0
    Novel3.0
    AI first pass
  12. 12
    OpenAI opens its Decisions API in public beta, built on GPT-6 Luna
    developers.openai.com§ Decisions in the wildOpenAI6 Oct
    makerOpenAItasktyped decisionstaskrouting and classification
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful4.0
    Notable4.0
    Novel3.0
    AI first pass
  13. 13
    Torchcast AI releases Torchcast Decision 27B, built on StartLux weights
    huggingface.co§ Decisions in the wildTorchcast AI (torchcast-ai)3 Oct
    makerTorchcast AItasktyped decisions
    Hype gap ▼ −0.8 under
    Shipped
    Real3.5
    Useful3.0
    Notable2.5
    Novel2.5
    AI first pass
  14. 14
    Aleph Alpha releases Kolibri 1, a 78B German and English reasoning MoE under Apache 2.0
    huggingface.co§ Open weightsAleph Alpha GmbH (Aleph-Alpha)3 Oct
    makerAleph Alpha GmbHtaskGerman-language assistantstaskretrieval-augmented generation
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  15. 15
    SberAI releases FRIDA-Decisions, a Russian and English encoder for structured decisions
    huggingface.co§ Decisions in the wildSberAI (ai-forever)2 Oct
    makerSberAItasktyped decisionstaskRussian-language classification
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel3.5
    AI first pass
  16. 16
    H2O.ai releases H2O Lightning 4B, an Apache-2.0 decision model with GGUF builds
    huggingface.co§ Decisions in the wildH2O.ai (h2oai)2 Oct
    makerH2O.aitasktyped decisions
    Hype gap ▼ −1.0 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel2.5
    AI first pass
  17. 17
    Perplexity releases pplx-decider, an Apache-2.0 27B decision model
    huggingface.co§ Decisions in the wildPerplexity (perplexity-ai)1 Oct
    makerPerplexitytasktyped decisionstaskclassification
    Hype gap ▼ −0.2 under
    Shipped
    Real3.5
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  18. 18
    Fish Audio releases Drama 3, a 35B mixture-of-experts TTS for multi-character scenes
    huggingface.co§ Open weightsFish Audio (fishaudio)1 Oct
    makerFish Audiotaskmulti-speaker speechtaskvoice cloning
    Hype gap ▬ 0.0 level
    Shipped
    Real3.5
    Useful3.5
    Notable3.5
    Novel3.5
    AI first pass
  19. 19
    vLLM Semantic Router team releases Decision 2.0 open decision models
    huggingface.co§ Decisions in the wildvLLM Semantic Router (vllm-sr)29 Sep
    makervLLM Semantic Routertaskrequest routingtasktyped decisions
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  20. 20
    An independent developer releases Jev-Style 2B Decision v3, a 2B model with a free hosted API
    huggingface.co§ Decisions in the wildchaoliang yan (chaoliangUNSW)27 Sep
    makerchaoliang yantasktyped decisions
    Hype gap ▼ −0.8 under
    Shipped
    Real3.5
    Useful2.5
    Notable2.0
    Novel2.5
    AI first pass
  21. 21
    Shanghai AI Laboratory releases Intern-Decision, multimodal decision models from 0.8B to 4B
    huggingface.co§ Decisions in the wildShanghai AI Laboratory (internlm)26 Sep
    makerShanghai AI Laboratorytasktyped decisionstaskmultimodal classification
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel3.0
    AI first pass
  22. 22
    Xiaomi MiMo releases MiMo-V2.6-Pro-RL, a 1.02T omnimodal MoE under MIT
    huggingface.co§ Open weightsXiaomi MiMo (XiaomiMiMo)21 Sep
    makerXiaomi MiMotaskcoding agentstaskreinforcement learning
    Hype gap ▬ 0.0 level
    Shipped
    Real3.5
    Useful3.5
    Notable3.5
    Novel3.5
    AI first pass
  23. 23
    distil labs releases an invoice-decision model that approves or holds vendor invoices
    huggingface.co§ Decisions in the wilddistil labs (distil-labs)21 Sep
    makerdistil labstaskinvoice approvaltaskthree-way match
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.0
    Notable2.5
    Novel3.0
    AI first pass
  24. 24
    JetBrains releases Mellum2.1 Thinking, a 12B coding MoE with 2.5B active, under Apache 2.0
    huggingface.co§ Open weightsJetBrains (JetBrains)20 Sep
    makerJetBrainstaskcodingtasksoftware engineering agents
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel3.0
    AI first pass
  25. 25
    Jared Palmer releases Kev-9B, an Apache-2.0 LoRA decision model on Qwen3.5-9B
    huggingface.co§ Decisions in the wildJared Palmer (jaredpalmer)20 Sep
    makerJared Palmertasktyped decisionstaskdocument classification
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel3.0
    AI first pass
  26. 26
    Convai Innovations releases laya-typed-decisions, a 0.4B ModernBERT decision model
    huggingface.co§ Decisions in the wildConvai Innovations (convaiinnovations)18 Sep
    makerConvai Innovationstasktyped decisions
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.0
    Notable2.5
    Novel3.0
    AI first pass
  27. 27
    TypeSafe AI launches Jev, a hosted model for typed software decisions
    typesafe.ai§ Decisions in the wildTypeSafe AI (typesafe-ai)15 Sep
    makerTypeSafe AItasktyped decisionstaskrouting and classification
    Hype gap ▲ +0.8 over
    Pending
    Real3.0
    Useful4.0
    Notable4.5
    Novel4.0
    AI first pass
  28. 28
    Yandex releases AliceAI-Foundation-80B-A3B-Base, an Apache 2.0 base MoE for Russian and English
    huggingface.co§ Open weightsyandex (yandex)12 Sep
    makeryandextaskRussian-language modellingtaskfine-tuning base
    Hype gap ▬ 0.0 level
    Shipped
    Real4.0
    Useful3.0
    Notable3.5
    Novel3.5
    AI first pass
  29. 29
    DeepSeek releases V4.1-Flash, a multimodal MoE with a much smaller KV cache, under MIT
    huggingface.co§ Open weightsDeepSeek-AI (deepseek-ai)10 Sep
    makerDeepSeek-AItasklong-context inferencetaskagentic workloads
    Hype gap ▲ +0.2 over
    Shipped
    Real4.0
    Useful3.5
    Notable4.0
    Novel4.0
    AI first pass
  30. 30
    OpenBMB releases MiniCPM5-2B, a dense 2B model with open training data, under Apache 2.0
    huggingface.co§ Open weightsOpenBMB (openbmb)6 Sep
    makerOpenBMBtaskon-device inferencetasktool use
    Hype gap ▼ −0.8 under
    Shipped
    Real4.5
    Useful3.5
    Notable4.0
    Novel2.5
    AI first pass
  31. 31
    Z.ai releases GLM-5.3, a 753B open-weights model aimed at coding and cyber tasks
    huggingface.co§ Open weightsZ.ai (zai-org)25 Aug
    makerZ.aitaskcodingtaskvulnerability discovery
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful4.0
    Notable4.0
    Novel2.5
    AI first pass
  32. 32
    BreezeBlue releases Breeze-TTS-2, a 3B English and Chinese TTS, non-commercial
    huggingface.co§ Open weightsBreezeBlue (BreezeBlue)25 Aug
    makerBreezeBluetaskvoice cloningtaskvoice design
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  33. 33
    Qwen releases Qwen3.8-27B, a dense 27B vision-language model under Apache 2.0
    huggingface.co§ Open weightsQwen (Qwen)5 Aug
    makerQwentaskcodingtaskagentic tool use
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful4.0
    Notable4.5
    Novel3.0
    AI first pass
  34. 34
    NVIDIA releases MagpieTTS Multilingual 357M v2607, 12 languages with preset voices
    huggingface.co§ Open weightsNVIDIA (nvidia)21 Jul
    makerNVIDIAtaskmultilingual speech
    Hype gap ▼ −1.0 under
    Shipped
    Real4.5
    Useful3.5
    Notable3.5
    Novel2.5
    AI first pass
  35. 35
    Mistral AI releases Shieldstral 1.0, a 3B multimodal safety classifier under Apache 2.0
    huggingface.co§ Decisions in the wildMistral AI (mistralai)16 Jul
    makerMistral AItaskcontent moderationtaskmultimodal safety classification
    Hype gap ▼ −0.8 under
    Shipped
    Real4.5
    Useful4.0
    Notable3.5
    Novel3.5
    AI first pass
  36. 36
    Moonshot releases Kimi K3, a 2.8T-parameter open-weights multimodal model
    huggingface.co§ Open weightsmoonshotai (moonshotai)13 Jun
    makermoonshotaitasklong-horizon codingtaskagentic knowledge work
    Hype gap ▲ +0.8 over
    Shipped
    Real3.5
    Useful4.0
    Notable5.0
    Novel4.0
    AI first pass
  37. 37
    Boson AI releases Higgs TTS 3, a 4B zero-shot voice cloning model for 100+ languages
    huggingface.co§ Open weightsBoson AI (bosonai)4 Jun
    makerBoson AItaskvoice cloningtaskmultilingual speech
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.5
    Notable4.0
    Novel3.0
    AI first pass
  38. 38
    NVIDIA releases Nemotron 3.5 Content Safety, a 4B multilingual moderation classifier
    huggingface.co§ Decisions in the wildNVIDIA (nvidia)2 Jun
    makerNVIDIAtaskcontent moderation
    Hype gap ▼ −1.0 under
    Shipped
    Real4.5
    Useful4.0
    Notable3.5
    Novel3.0
    AI first pass
  39. 39
    StepFun releases Step 3.7 Flash, a 198B vision-language MoE with 11B active parameters
    huggingface.co§ Open weightsStepFun (stepfun-ai)23 May
    makerStepFuntaskagentic workflowstaskvisual reasoning
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  40. 40
    Supertone releases Supertonic 3, a 99M on-device TTS in 31 languages
    huggingface.co§ Open weightsSupertone (Supertone)6 May
    makerSupertonetaskon-device speech synthesis
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  41. 41
    IBM releases Granite 4.1 8B, an Apache 2.0 instruct model for enterprise tasks
    huggingface.co§ Open weightsIBM (ibm-granite)29 Apr
    makerIBMtaskretrieval-augmented generationtasktext extraction
    Hype gap ▼ −1.2 under
    Shipped
    Real4.5
    Useful3.5
    Notable3.0
    Novel2.5
    AI first pass
  42. 42
    IBM releases Granite Guardian 4.1 8B, a judge model for safety, RAG and agent checks
    huggingface.co§ Decisions in the wildIBM (ibm-granite)16 Apr
    makerIBMtaskcontent moderationtaskLLM-as-judge
    Hype gap ▼ −1.0 under
    Shipped
    Real4.5
    Useful4.0
    Notable3.5
    Novel3.0
    AI first pass
  43. 43
    OpenBMB releases VoxCPM2, a 2B Apache 2.0 TTS with voice design and 48 kHz output
    huggingface.co§ Open weightsOpenBMB (openbmb)3 Apr
    makerOpenBMBtaskvoice cloningtaskvoice design
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful4.0
    Notable4.0
    Novel3.5
    AI first pass
  44. 44
    Mistral releases Medium 3.5, a dense 128B model with a modified MIT licence
    huggingface.co§ Open weightsMistral AI (mistralai)31 Mar
    makerMistral AItaskcodingtaskagentic function calling
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful4.0
    Notable3.5
    Novel3.0
    AI first pass
  45. 45
    k2-fsa releases OmniVoice, a 0.6B zero-shot TTS covering 600+ languages
    huggingface.co§ Open weightsk2-fsa (k2-fsa)30 Mar
    makerk2-fsataskvoice cloningtaskmultilingual speech
    Hype gap ▬ 0.0 level
    Shipped
    Real4.0
    Useful3.5
    Notable4.0
    Novel3.5
    AI first pass
  46. 46
    Mistral releases Voxtral TTS, a 4B open-weights speech model, non-commercial
    huggingface.co§ Open weightsMistral AI (mistralai)26 Mar
    makerMistral AItaskvoice agentstaskmultilingual speech
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.5
    Notable4.0
    Novel3.0
    AI first pass
  47. 47
    NVIDIA releases Nemotron 3 Content Safety, a 4B moderation classifier on Gemma-3
    huggingface.co§ Decisions in the wildNVIDIA (nvidia)16 Mar
    makerNVIDIAtaskcontent moderation
    Hype gap ▼ −1.2 under
    Shipped
    Real4.5
    Useful3.5
    Notable3.0
    Novel2.5
    AI first pass
  48. 48
    Google DeepMind releases Gemma 4, with a 31B dense model under Apache 2.0
    huggingface.co§ Open weightsGoogle DeepMind (google)11 Mar
    makerGoogle DeepMindtaskon-device inferencetaskmultimodal reasoning
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful4.0
    Notable4.5
    Novel3.0
    AI first pass
  49. 49
    Fish Audio releases S2 Pro, a 5B TTS model for 80+ languages, research licence
    huggingface.co§ Open weightsFish Audio (fishaudio)9 Mar
    makerFish Audiotaskexpressive speechtaskmultilingual speech
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  50. 50
    NineNineSix releases KaniTTS2-en, a 400M English TTS with voice cloning
    huggingface.co§ Open weightsNineNineSix (nineninesix)12 Feb
    makerNineNineSixtaskvoice cloningtaskreal-time speech
    Hype gap ▼ −0.8 under
    Shipped
    Real4.0
    Useful3.0
    Notable2.5
    Novel3.0
    AI first pass
  51. 51
    OpenMOSS Team releases MOSS-TTS, an 8B Apache 2.0 voice cloning and long-form TTS
    huggingface.co§ Open weightsOpenMOSS Team (OpenMOSS-Team)6 Feb
    makerOpenMOSS Teamtaskvoice cloningtasklong-form speech
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful4.0
    Notable3.5
    Novel3.5
    AI first pass
  52. 52
    Qwen releases Qwen3-TTS, Apache 2.0 voice cloning in 10 languages
    huggingface.co§ Open weightsQwen (Qwen)22 Jan
    makerQwentaskvoice cloningtasktext-to-speechtaskmultilingual speech
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful4.0
    Notable4.5
    Novel3.0
    AI first pass
  53. 53
    Hume AI releases TADA-1B, an English TTS that aligns one speech vector per text token
    huggingface.co§ Open weightsHume AI (HumeAI)12 Jan
    makerHume AItasktext-to-speech
    Hype gap ▲ +0.2 over
    Shipped
    Real3.5
    Useful3.0
    Notable3.0
    Novel4.0
    AI first pass
  54. 54
    Soprano-80M: an 80M Apache 2.0 TTS that claims 2000x real time on GPU
    huggingface.co§ Open weightsEugene Kwek (ekwek)22 Dec
    makerEugene Kwektaskon-device speech synthesis
    Hype gap ▼ −0.8 under
    Shipped
    Real4.5
    Useful3.0
    Notable3.0
    Novel3.0
    AI first pass
  55. 55
    FunAudioLLM releases Fun-CosyVoice3-0.5B-2512, an Apache 2.0 voice cloning TTS
    huggingface.co§ Open weightsFunAudioLLM (FunAudioLLM)11 Dec
    makerFunAudioLLMtaskvoice cloningtaskmultilingual speech
    Hype gap ▼ −1.0 under
    Shipped
    Real4.5
    Useful4.0
    Notable3.5
    Novel3.0
    AI first pass
  56. 56
    Z.ai releases GLM-TTS, an MIT-licensed Chinese and English voice cloning model
    huggingface.co§ Open weightsZ.ai (zai-org)10 Dec
    makerZ.aitaskvoice cloningtasktext-to-speech
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  57. 57
    NVIDIA releases Nemotron 3 Nano 30B-A3B, a hybrid Mamba-MoE reasoning model
    huggingface.co§ Open weightsNVIDIA (nvidia)4 Dec
    makerNVIDIAtaskreasoningtaskagentic coding
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful4.0
    Notable4.0
    Novel3.5
    AI first pass
  58. 58
    Microsoft releases VibeVoice-Realtime-0.5B, a streaming TTS model under MIT
    huggingface.co§ Open weightsMicrosoft (microsoft)4 Dec
    makerMicrosofttaskstreaming text-to-speech
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.0
    Notable3.5
    Novel3.0
    AI first pass
  59. 59
    Resemble AI releases Chatterbox-Turbo, a 350M MIT TTS built for voice agents
    huggingface.co§ Open weightsResemble AI (ResembleAI)2 Dec
    makerResemble AItaskvoice cloningtaskvoice agents
    Hype gap ▼ −0.8 under
    Shipped
    Real4.5
    Useful3.5
    Notable3.5
    Novel3.0
    AI first pass
  60. 60
    Ai2 releases Olmo 3, with a fully documented 32B Think model under Apache 2.0
    huggingface.co§ Open weightsAi2 (allenai)19 Nov
    makerAi2taskreasoning researchtaskmath and code
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful3.5
    Notable4.0
    Novel3.0
    AI first pass
  61. 61
    Nari Labs releases Dia2-2B, a streaming English dialogue TTS under Apache 2.0
    huggingface.co§ Open weightsNari Labs (nari-labs)15 Nov
    makerNari Labstaskdialogue speechtaskstreaming text-to-speech
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.0
    Notable3.0
    Novel3.5
    AI first pass
  62. 62
    StepFun releases Step-Audio-EditX, a 3B open-weights model for editing and generating speech
    huggingface.co§ Open weightsStepFun (stepfun-ai)12 Nov
    makerStepFuntasktext-to-speechtaskspeech editing
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful3.5
    Notable3.5
    Novel3.5
    AI first pass
  63. 63
    Moonshot releases Kimi K2 Thinking, a 1T MoE thinking agent with native INT4
    huggingface.co§ Open weightsmoonshotai (moonshotai)4 Nov
    makermoonshotaitaskagentic reasoningtasktool use
    Hype gap ▲ +0.2 over
    Shipped
    Real3.5
    Useful3.5
    Notable4.0
    Novel3.5
    AI first pass
  64. 64
    Mistral releases Ministral 3, edge models in 3B, 8B and 14B under Apache 2.0
    huggingface.co§ Open weightsMistral AI (mistralai)31 Oct
    makerMistral AItaskedge deploymenttaskimage understanding
    Hype gap ▼ −1.0 under
    Shipped
    Real4.5
    Useful3.5
    Notable3.5
    Novel2.5
    AI first pass
  65. 65
    OpenAI releases gpt-oss-safeguard, open-weight safety reasoning models that read your policy
    huggingface.co§ Decisions in the wildOpenAI (openai)29 Oct
    makerOpenAItaskcontent moderationtaskpolicy-based classification
    Hype gap ▼ −0.5 under
    Shipped
    Real4.5
    Useful4.0
    Notable4.0
    Novel3.5
    AI first pass
  66. 66
    Soul-AILab releases SoulX-Podcast-1.7B, a multi-speaker podcast TTS in English and Chinese
    huggingface.co§ Open weightsSoul-AILab (Soul-AILab)27 Oct
    makerSoul-AILabtaskmulti-speaker speechtaskvoice cloning
    Hype gap ▼ −0.2 under
    Shipped
    Real3.5
    Useful3.5
    Notable3.0
    Novel3.5
    AI first pass
  67. 67
    Qwen releases Qwen3Guard, multilingual moderation models in 0.6B, 4B and 8B
    huggingface.co§ Decisions in the wildQwen (Qwen)23 Sep
    makerQwentaskcontent moderation
    Hype gap ▼ −0.8 under
    Shipped
    Real4.5
    Useful4.0
    Notable3.5
    Novel3.5
    AI first pass
  68. 68
    Skywork releases Reward-V2, eight reward models from 0.6B to 8B
    huggingface.co§ Decisions in the wildSkywork (Skywork)2 Jul
    makerSkyworktaskreward modelling
    Hype gap ▼ −1.0 under
    Shipped
    Real4.5
    Useful4.0
    Notable3.5
    Novel3.0
    AI first pass
  69. 69
    Katanemo releases Arch-Router 1.5B, a router that follows your own route descriptions
    huggingface.co§ Decisions in the wildKatanemo (katanemo)19 Jun
    makerKatanemotaskLLM routing
    Hype gap ▼ −0.5 under
    Shipped
    Real4.0
    Useful3.5
    Notable3.0
    Novel3.5
    AI first pass
  70. 70
    Meta releases Llama Guard 4, a 12B multimodal safety classifier
    huggingface.co§ Decisions in the wildMeta (meta-llama)23 Apr
    makerMetataskcontent moderation
    Hype gap ▼ −0.2 under
    Shipped
    Real4.0
    Useful3.5
    Notable4.0
    Novel3.0
    AI first pass