Fireworks AI has announced that DeepSeek V4 Flash now includes native vision capabilities. This update is available on both serverless and priority tiers, offering improved benchmarks compared to the text-only 0731 version at a similar price point. The new vision feature allows the model to process visual input, enhancing its utility for various tasks. AI
IMPACT Enhances a widely-used model with multimodal capabilities, potentially broadening its applications in visual understanding tasks.
RANK_REASON Model release announcement from a frontier lab (DeepSeek) with system card. [lever_c_demoted from frontier_release: ic=2 ai=1.0]
Read on X — Fireworks (inference infra) →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →