new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

AI Asset Pipeline · all subjects

api/models

2 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Image generation model IDs and capabilities

Available models with aliases: nb2 (gemini-3.1-flash-image-preview, supports 0.5K–4K), nb (gemini-2.5-flash-image-preview), pro (gemini-3-pro-image-preview), gpt2 (gpt-image-2, best typography, n=1 only, supports WxH size format), gpt1 (gpt-image-1), imagen (imagen-4.0), seedream/seedream45 (doubao-seedream-4.5, n up to 15, 2K default), seedream40 (doubao-seedream-4-0), seedream5 (doubao-seedream-5.0-lite), qwen (qwen-image), zturbo (z-image-turbo, cheap/fast with low quality), grok (grok-imagine), wan (wan2.7-image). Gemini family has -official variants used when provider=official.

aimlapi sound generation models and cost per call

Sound generation models: stable-audio ($0.016 per call, use for all SFX—only model doing short audio), lyria2 ($0.013 per call, BGM default but flaky with 503s and fallthrough), minimax/music-2.6 ($0.195 per call, BGM workhorse with true is_instrumental), elevenlabs/eleven_music (~$3.29 per call, best quality but requires explicit opt-in only). Retrieve current model ids via `from soundgen.aimlapi import catalog; catalog()` and never guess ids.

Give your agent this brain