GROUNDED IN REAL HUMAN CONVERSATIONSYOUR BENCH
LAUNCHES

New AI models, last 30 days

Every model that appeared in the last 30 days, with how many people are talking about it and the community verdict so far.

NEW ARRIVALS

29 MODELS
  • Mercury 2.5InceptionTEXT62 voices#216

    Not enough discussion about Mercury 2.5 yet to call it. We found 62 posts but people did not say much either way.

  • Nex N2.5 Mini (free)Nex AgiMULTIMODAL20 voices#391

    Not enough discussion about Nex N2.5 Mini (free) yet to call it. We found 20 posts but people did not say much either way.

  • Nex N2.5 Pro (free)Nex AGIMULTIMODAL0 voices

    Nex-N2.5 is an agentic model built to turn goals into working, verified outcomes. Its core strength is agentic coding within a visual feedback loop: it can explore codebases, implement multi-file...

  • GPT 6 AstraOpenAIMULTIMODAL3,677 voices#197

    People are mostly frustrated with GPT 6 Astra right now. Biggest praise: speed. Biggest gripe: working when you need it.

  • GPT 6 Astra ProOpenAIMULTIMODAL1,258 voices#405

    People are mostly frustrated with GPT 6 Astra Pro right now. Biggest gripe: working when you need it.

  • Ling 3.0 Flash Sante (free)InclusionaiTEXT0 voices

    Ling 3.0 Flash Sante is a health and medicine-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for...

  • Muse Spark 1.3MetaTEXT658 voices#43

    People generally like Muse Spark 1.3, with some gripes. Biggest praise: speed. Biggest gripe: working when you need it.

  • Muse Spark 1.3 ContributorMetaMULTIMODAL4 voices

    Meta's Muse Spark 1.3 Contributor is a frontier multimodal reasoning model for long-horizon coding and agentic workflows, with strong gains in computer use, browsing, professional tool use, codebase understanding, and million-token retrieva

  • Gemini 3.8 FlashGoogleMULTIMODAL1,371 voices#78

    People generally like Gemini 3.8 Flash, with some gripes. Biggest praise: help with code. Biggest gripe: saying no too much.

  • Claude Fable 5.1AnthropicMULTIMODAL11,948 voices#170

    The internet is split on Claude Fable 5.1. Biggest praise: help with code. Biggest gripe: working when you need it.

  • Granite 4.2 8BIbm GraniteTEXT71 voices#425

    Not enough discussion about Granite 4.2 8B yet to call it. We found 71 posts but people did not say much either way.

  • Ling 3.0 Flash FinInclusionaiTEXT0 voices

    Ling 3.0 Flash Fin is a finance-focused mixture-of-experts model from InclusionAI, built on Ling 3.0 Flash with 5.1B active parameters out of 124B total. It is designed for real-world investment...

  • Qwen3.8 FlashAlibabaMULTIMODAL5,155 voices#5

    The internet is split on Qwen3.8 Flash. Biggest praise: writing. Biggest gripe: saying no too much.

  • GLM 5.3 FlashZhipuMULTIMODAL6,884 voices#10

    The internet is split on GLM 5.3 Flash. Biggest praise: help with code. Biggest gripe: saying no too much.

  • Muse Spark 1.2 ContributorMetaTEXT0 voices

    Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...

  • DeepSeek V4 Flash Vision ExpDeepSeekMULTIMODAL4,368 voices#18

    People are mostly happy with DeepSeek V4 Flash Vision Exp. Biggest praise: price.

  • Hy MT2 1.8BTencentTEXT798 voices#39

    The internet is split on Hy MT2 1.8B. Biggest praise: writing. Biggest gripe: getting facts right.

  • Hy MT2 30B A3BTencentTEXT2 voices

    Hy-MT2-30B-A3B is Tencent's flagship translation model in the Hy-MT2 family. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and

  • Hy MT2 7BTencentTEXT0 voices

    Hy-MT2-7B is a 7B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided tra

  • GLM 5.3ZhipuMULTIMODAL9,861 voices#4

    People generally like GLM 5.3, with some gripes. Biggest praise: help with code. Biggest gripe: working when you need it.

  • Qwen3.8 27BAlibabaMULTIMODAL14,007 voices#2

    The internet is split on Qwen3.8 27B. Biggest praise: help with code. Biggest gripe: saying no too much.

  • Dots3 NoteXiaohongshuMULTIMODAL2,634 voices#14

    People generally like Dots3 Note, with some gripes. Biggest praise: help with code. Biggest gripe: saying no too much.

  • Gemini 3.7 FlashGoogleMULTIMODAL6,815 voices#23

    People generally like Gemini 3.7 Flash, with some gripes. Biggest praise: help with code. Biggest gripe: working when you need it.

  • Seed 2.1 TurboBytedance SeedTEXT21 voices

    Seed 2.1 Turbo is a multimodal model from ByteDance Seed for coding and long-horizon agent workflows. It is suited for end-to-end software delivery, multi-step task execution, and understanding visual and...

  • Qwen3.8 2.4T A95BAlibabaMULTIMODAL1,023 voices#29

    People generally like Qwen3.8 2.4T A95B, with some gripes. Biggest praise: help with code. Biggest gripe: saying no too much.

  • Seed 2.0 CodeBytedance SeedTEXT0 voices

    Seed 2.0 Code is a model from ByteDance Seed optimized for agentic coding. It is suited for frontend development, multilingual programming tasks, and coding-agent workflows in tools such as Claude...

  • Grok 4.6xAIMULTIMODAL42,295 voices#1

    People generally like Grok 4.6, with some gripes. Biggest praise: help with code. Biggest gripe: working when you need it.

  • LFM2.5 2.6B (free)LiquidTEXT39 voices#264

    People are mostly happy with LFM2.5 2.6B (free). Biggest praise: working when you need it.

  • Sakana NamazuSakanaMULTIMODAL3 voices

    Sakana Namazu is a Japanese-specialized reasoning model from Sakana AI, based on Kimi K2.6 with additional training for Japanese language and business contexts. It is suited for Japanese instruction following,...