β¨ Intern-AI, Free sign-up, no card needed, no catch found yet.
β Details:
πΊ ~100M tokens (10 points, 1 point ~= 20M tokens)
πΊ Models: GLM-5.3, Qwen3.8-27B, Atria Dawn, Kimi-2.6, DeepSeek V4 (Pro/Flash)
πΊ ~80 hours of compute onNvidia A100-1-80G or Ascend910B-1-64G
πΊ 5-hour and weekly usage limits per model, standard for their tier
π₯ GPU on rent costs $1-2/hour means that's $80-160 of compute plus the full token stack, free.
π Sign up at discovery.intern-ai.org.cn
π Credits: @Atenov_D (on X)
π @BotRoomz














