DeepMind Releases Gemma 4 Vision-Capable Models
Google DeepMind on April 2, 2026 released Gemma 4, four vision-capable Apache 2.0–licensed reasoning LLMs sized 2B, 4B, 31B and a 26B-A4B mixture-of-experts. The models introduce Per-Layer Embeddings to boost effective parameter efficiency for on-device use; Google provides API access to the larger models via AI Studio while community GGUF builds enable local testing.
Scoring Rationale
Official same-day release from DeepMind with novel efficiency technique and open-source licensing drives high novelty, scope, and actionability; immediate local GGUF usability and API access justify top score despite a reported 31B instability.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Logistics & Shipping problemsStep-by-step roadmaps from zero to job-ready — curated courses, salary data, and the exact learning order that gets you hired.
Sources
- Read OriginalGemma 4: Byte for byte, the most capable open modelssimonwillison.net
