DeepSeek Launches V4-Flash-Vision-Exp, Nearing Opus 4.8 on Multimodal Agent Benchmarks
On August 21, DeepSeek launched the experimental multimodal model DeepSeek-V4-Flash-Vision-Exp on its official API platform. The model matches V4-Flash in text capabilities, covering agents, reasoning, and world knowledge; it shows a major leap over V4-Flash on multimodal agent benchmarks, approaching Opus 4.8 — well worth watching.
Confirmed
- The model is now live on the DeepSeek API platform as an experimental release (announced via the official @deepseekai account).
- Its text-only capabilities are on par with the official V4-Flash.
- Vision-related agent benchmark performance has improved substantially, approaching Opus 4.8.
- Officials note the model works smoothly across various agent frameworks, combining visual understanding with diverse tools to build multimodal agent workflows.
- Third-party user @TheCryptoCat75 mentioned V4 Flash can be accessed via DeepInfra at a lower price than the official API without peak-hour rate limits (a personal user experience).
Why it matters
- This release marks DeepSeek's push into multimodal agent capabilities, with a Flash-tier model approaching the top closed-source Opus 4.8 on multimodal agent benchmarks — a clear cost-efficiency path.
- Cross-agent-framework compatibility means developers can plug vision capabilities directly into existing workflows, lowering the barrier for multimodal applications.
2026-08-21 ~ 2026-08-21 · 9 related posts
Primary sources
- DeepSeek Launches V4-Flash Multimodal Model, Agent Capability Near Opus-4.8 — DeepSeek · 2026-08-21
- [source] DeepSeek launches V4-Flash-Vision-Exp, multimodal agent performance nears Opus 4.8 — deepseek_ai · 2026-08-21
- [source] DeepSeek: V4-Flash-Vision-Exp works smoothly across agent frameworks — deepseek_ai · 2026-08-21
- [source] DeepSeek Vision API details: up to 384 tokens per image at V4-Flash pricing — deepseek_ai · 2026-08-21
- DeepSeek V4-vision-exp launches API with ultra-fast speed and low cost — teortaxesTex · 2026-08-21
- Deepseek Launches Multimodal Model V4FlashVisionEXP — 歸藏的AI工具箱 · 2026-08-21
- DeepSeek V4 Flash adds vision capabilities, now live — TheCryptoCat75 · 2026-08-21
2 near-duplicate retellings: teortaxesTex · teortaxesTex