Qwen3.8-Flash slashes prices: flat 2.7 RMB output, cheapest cache hits yet
AlchainHust · x · 2026-08-27
Alibaba released Qwen3.8-Flash along with an open-weights next-gen architecture version. API pricing is a flat 0.8 RMB/M input tokens, 2.7 RMB/M output, 0.1 RMB cache hits — no peak/off-peak tiers, vs DeepSeek-V4-Flash's 4.5–9 RMB output. 100M input tokens cost 80 RMB vs $500 on Claude Opus 4.6.
Official benchmarks compare against Qwen3.8-27B, Qwen3.7-Plus, DeepSeek-V4-Flash and Claude Opus 4.6. The author also cites a V2EX post where a company moved to shift-based scheduling to dodge peak-hour token pricing.
More from Models
- China's StepFun releases Step 5 Preview weights: 600B MoE, 27B active, 1M context — Dr_Singularity · 2026-10-15
- StepFun open-sources Step 5 Preview: 600B MoE with 1M context, No. 2 on OpenRouter — Dr_Singularity · 2026-10-15
- Google says teams across the company are using Gemini 4 Argon on its hardest engineering problems — CurieuxExplorer · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Leak: Moonshot's Kimi flagship 'x4' done with pre-training, K3.1 in internal testing — lyraxana · 2026-10-15