Topic: model performance benchmarks

  • DeepSeek launches V4-Pro and V4-Flash, one year after historic breakthrough

    DeepSeek launches V4-Pro and V4-Flash, one year after historic breakthrough

    DeepSeek released preview versions of DeepSeek-V4-Pro and DeepSeek-V4-Flash, positioning them as the most powerful open-source AI models, directly challenging OpenAI and Anthropic. The V4 models feature a Hybrid Attention Architecture with a 1-million-token context window, designed for agentic an...

    Read More »
  • India's Murf AI launches Falcon 2 to rival voice AI giants

    India's Murf AI launches Falcon 2 to rival voice AI giants

    Murf AI will launch its Falcon 2 voice model on August 20 at $0.01 per generated minute, roughly one-fifth the cost of rival ElevenLabs, making the price gap a major selling point for high-volume enterprise users. Falcon 2 claims superior performance on the Artificial Analysis benchmark, beating ...

    Read More »
  • Mixed Reactions to OpenAI's New Open Source GPT Models

    Mixed Reactions to OpenAI's New Open Source GPT Models

    OpenAI released its first major open-source models since 2019, gpt-oss-120B and gpt-oss-20B, sparking debate over their competitive performance in benchmarks but limitations in real-world use. The Apache 2.0-licensed models allow local AI deployment, with gpt-oss-120B for enterprise h...

    Read More »
  • Alibaba challenges US AI dominance with new push

    Alibaba challenges US AI dominance with new push

    Alibaba released Qwen3.8-Max, its largest and most advanced AI model with 2.4 trillion parameters, which benchmarks show rivals or surpasses Anthropic's Fable 5 on several tests. The company will open-weight the model next week, a move that contrasts with closed US rivals like OpenAI and Anthropi...

    Read More »
  • DeepSeek previews new AI model closing gap with frontier systems

    DeepSeek previews new AI model closing gap with frontier systems

    DeepSeek has quietly released two preview versions of its next-generation model, DeepSeek V4 (Flash and Pro), built on a mixture-of-experts architecture with a 1 million token context window, making V4 Pro the largest open-weight model ever at 1.6 trillion total parameters. DeepSeek claims the V4...

    Read More »
  • Time Magazine Leaks Embargoed Story on New Anthropic AI Model

    Time Magazine Leaks Embargoed Story on New Anthropic AI Model

    Anthropic's upcoming AI model announcement was accidentally leaked by Time Magazine, revealing potential names like "Neptune" or "Claude Opus" for a next-generation Claude model. The leaked details suggest the new model will be a major upgrade with expanded parameters, offering greater complexity...

    Read More »