DeepSeek launches V4-Flash-Vision-Exp, multimodal agent performance nears Opus 4.8
deepseek_ai · x · 2026-08-21
DeepSeek has launched the experimental multimodal model DeepSeek-V4-Flash-Vision-Exp on its API platform.
- Text capabilities match V4-Flash, covering agents, reasoning, and world knowledge.
- On multimodal agent benchmarks it makes a major leap over V4-Flash, bringing performance close to Anthropic's Opus 4.8.
- Available via model='deepseek-v4-flash-vision-exp'; DeepSeek Harness 0.1.1 released the same day with out-of-the-box support.
Related event: DeepSeek launches experimental multimodal model V4-Flash-Vision-Exp(5 posts)→
More from Models
- LLM German Output Cringed: Reads Like It Was Written by Olaf Scholz — DominiqueCAPaul · 2026-08-21
- DeepSeek V4-vision-exp launches API with ultra-fast speed and low cost — teortaxesTex · 2026-08-21
- Fastest NVFP4 quant of Qwen3.8 27B released, 50% faster than Q4 on compatible hardware — ionsago · 2026-08-21
- Jie Tang on scaling history: FLOPs were intelligence, parameters were knowledge — cedric_chee · 2026-08-21
- Leaked Benchmark Suggests MIMO V3 Pro Performance Rivals Fable — teortaxesTex · 2026-08-21
- Leaked GLM-5.4 Reportedly Surpasses GPT Thanks to Rapid RL Iteration — kimmonismus · 2026-08-21