DEEPSEEK HAS SHIPPED SINCE — CHECKED 8 OCTOBER 2026
DeepSeek V4.1-Flash was released on 10 September 2026 and replaces the Flash tier compared below.
● Price: $0.30 input / $1.20 output per million tokens at peak hours
● Licence: MIT
● V4-Pro remains a current product and is unaffected
The cost reasoning below still applies — what changed is which Flash model you would actually call.
QUICK VERDICT — AUGUST 2026
V4 Flash 0731 for agentic coding: $0.14/M, Terminal-Bench 82.7%, Codex-compatible, MIT weights. The verified cost-performance leader for agentic coding workloads today. Migrate from deepseek-chat before October 24, 2026.
Qwen3.8-Max-Preview for multimodal evaluation: Test via Token Plan for document-heavy, image, or general frontier tasks. Not ready for production until Alibaba publishes benchmarks, per-token pricing, and open-weight checkpoint.
Last updated August 3, 2026. Related: Qwen3.8-Max full fact-check → · DeepSeek V4 Flash 0731 review →