DeepSeek V4.1 Flash vs Qwen 3.8 Flash
قارن النماذج
Qwen 3.8 Flash is more worth it — 74% vs 54%
How we score worth-it? →DeepSeek's open 485B-parameter Flash variant of V4.1.
- Insane improvement in kvcache compression. Now 1M context is only ~900MB.— reddit ↗
- Benchmarks are good! A bit below frontier in general knowledge but almost on par with Sol and Opus in agentic coding.— reddit ↗
- And most importantly: on their API it's 100x cheaper, 10-20x faster.— reddit ↗
- It's flash because it's fast. It uses something ridiculous like 9B parameters for prefilling.— reddit ↗
- I will continue to be amazed by how much power you get from DeepSeek Flash for the cost. I have let that puppy loose on so many projects and it has never let me down.— hn ↗
- Looks like even 4x H20 can't run FP8 anymore.— v2ex ↗
- Sadly, it also tops the hallucination benchmark. They really have to improve the hallucination rate.— reddit ↗
- What's concerning is that they called it 'Flash'. If this is a new 'Flash' standard, then the happy days of running DeepSeek locally are over.— reddit ↗
- First time I see DeepSeek offering multiple options, asking obnoxious questions and digging into intentions, while providing 'insight' that reads more like rambling than actual advice.— reddit ↗
- The web UI chat version of flash has very poor language following abilities: ask in English and get a thinking chain in Chinese with an answer in Chinese.— hn ↗
“Way too fast”
“485B model with a 510GB file — hard to run even quantized”
“The per-token KV cache is only 1/5 of before”
Alibaba's fast, cheap open-weight multimodal model (Qwen 3.8 family) with 1M context, deep thinking, and agentic coding — the budget Flash sibling to Qwen 3.8 Max.
- Qwen 3.8 Flash Next has only 6B activated parameters — on a single PRO 6000 it pre-fills over 10K tokens and decodes at 150 tps, with room to optimize. It really is sparse and fast.— v2ex ↗
- Qwen 3.8 flash next is way better than 27B. It's so good I don't even use Claude anymore.— hn
- Qwen Flash Next not only seemed less delusional than the 27B, but also made useful observations and problem-solved ways around many website access issues.— hn ↗
- Qwen3.8-next-flash is impressive though, especially with Strata or Basalt as inference engines. On a 5090 laptop with 64GB RAM one can run NVFP4 quants at 120-140t/s decode and 3000-4000t/s prefill.— hn ↗
- Qwen 3.8's Flash Next at A6B seems about juuuuust right— hn ↗
“Sparse and fast — only 6B active parameters, decoding at 150 tps on a single consumer GPU.”
“Flattened the VRAM barrier — a data-center-grade long-context model that runs on a consumer RTX 4090.”
“Superior quality for coding, and notably good at finding issues in generated code.”
Which should you pick?
The gap is clear: Qwen 3.8 Flash sits at 74% while DeepSeek V4.1 Flash is at 54%, a 20-point difference. If you just want the safer pick, start with Qwen 3.8 Flash.
On entry price, Qwen 3.8 Flash is cheaper ($0.11 vs $0.15, DeepSeek V4.1 Flash costs about 36% more).
Both are in the محادثة category with similar coverage; the real difference is stability and cost — run one of your own tasks through each before deciding.
What users actually said about Qwen 3.8 Flash: "Qwen 3.8 Flash Next has only 6B activated parameters — on a single PRO 6000 it pre-fills o…"
FAQ
Is DeepSeek V4.1 Flash or Qwen 3.8 Flash more worth it?
Qwen 3.8 Flash is more worth it, with a 74% worth-it score vs 54% for DeepSeek V4.1 Flash.
How much does DeepSeek V4.1 Flash cost vs Qwen 3.8 Flash?
DeepSeek V4.1 Flash: $0.15/$0.60 per 1M · Qwen 3.8 Flash: $0.11 / $0.38 per 1M tokens.
What do users say about DeepSeek V4.1 Flash vs Qwen 3.8 Flash?
DeepSeek V4.1 Flash: Way too fast. Qwen 3.8 Flash: Sparse and fast — only 6B active parameters, decoding at 150 tps on a single consumer GPU..