DeepSeek debuts V4-Flash AI model at $0.03 per test
DeepSeek, a Chinese startup, just dropped its new V4-Flash AI model, and it's making waves for being the most affordable out there.
Running V4-Flash costs only $0.14 per million input tokens and $0.28 per million output tokens, which averages out to about 3 cents per benchmark test.
For comparison, that's more than 100 times cheaper than Anthropic's Claude Fable 5 ($3.15 per test).
V4-Flash scores 50 on intelligence index
While V4-Flash is super budget-friendly, its Intelligence Index score sits at 50, matching Google's Gemini 3.6 Flash but trailing Moonshot's Kimi K3 (57) and Meta's Muse Spark 1.1 (51).
DeepSeek is aiming big with plans for an IPO and work on a more advanced V4-Pro model (no release date yet), but it will have to compete with strong rivals like Moonshot, Alibaba, and global tech giants in the crowded AI market.