DeepSeek V4.1 Flash
Cheap, fast, and good enough for most of the work I used to think needed a big model.
Tested 2026-11-08
Flash is the tier below Pro at a quarter of the price: $0.15 per million input tokens and $0.60 per million output. Cached input is $0.003 per million, which is close enough to free that I stop thinking about context size.
Most jobs are not hard. A script that reads a JSON file, a CSS change, a build script edit, a first pass at copy. Flash does those and I keep the expensive model for the jobs that actually need it. This is the model my routine work runs on.
Pros
- About 4.4x cheaper input and 3.3x cheaper output than V4 Pro
- Fast enough to sit in a loop without me noticing the wait
- Cached input at $0.003/M is effectively free
- Handles single-file work and small features without hand-holding
Cons
- Loses the thread on jobs that span many files and many turns
- More likely to make something up rather than stop and say it is unsure
- Same network and paid-key requirement as every cloud model
What I tested
| Test | Result |
|---|---|
| Long document summary | TODO: run the same test doc as V4 Pro so the two are comparable. |
| Coding task | Fine on one file. When the same task touched a template, a stylesheet and a build script, it lost one of the three changes and reported success anyway. |
| Hard reasoning | TODO: record the question and result. |
| Image understanding | TODO: not tested yet. |
| Cost for a week of real use | Priced at $0.15/M in, $0.60/M out, $0.003/M cached in. A 100K-token prompt with a 5K-token answer costs about $0.018. |
Bottom line: My default for everyday work. Switch up to V4 Pro the moment a task crosses a file boundary or goes live.
What this site is
I test AI tools myself — cloud models, local models, and the apps that manage them. Reviews only appear once I’ve run the same prompts through each tool on my own machine or account. No sponsorships, no affiliate links, no ads. More about how this works.
One email a week
What I tested this week, what it cost to run, and the one thing worth your time. No spam, unsubscribe in one click.