Jono Finger has used DeepSeek 4.1 Flash heavily across a dozen projects for a month. He says it often feels indistinguishable from pricier models in his own sessions. That’s a report about his workflow, not a head-to-head test: the interesting question is what becomes worth delegating when each experiment costs so little.

  • He uses the model for exploratory UI testing, file organization, planning and coding; he says his sessions rarely exceed $1 in expected cost, even when they run most of a day.
  • For some critical tasks he asks Opus 5.5 for a final code review, then has DeepSeek make the fixes. The stronger model supplies another set of eyes rather than doing every step.
  • Finger credits a much smaller key-value cache — the memory a model uses to keep track of a long conversation — for lower running costs. His further claims about water and electricity savings are plausible hypotheses, not measured comparisons in the essay.

His case is less “this model wins every comparison” than “cheap enough changes the workflow.” That distinction matters: being able to try more ideas is useful even when the model still needs checking.

The 199-comment thread on Hacker News adds firsthand counterexamples and a pricing wrinkle the essay’s own monthly bill cannot settle.

What the thread adds

  • p1necone — tried using Flash for both planning and implementation on a complex compiler project, then switched back to a stronger planner: it missed relevant context and introduced design decisions without discussion. They still find it “perfectly capable of being the sole agent for all of my well specced implementation tasks.”
  • gregwebs — reports spending $1–2 on an all-day session, but says Flash is “horrible at grilling sessions” for technical decisions. They use it for research and verification while testing stronger models for coding.
  • vishvananda — disputes the simple cheap-model comparison: heavily subsidized frontier subscriptions can cost less to a particular user than metered access to an open model. georgel replies with the opposite experience after months of coding through a different provider. These are individual bills, not a general price ranking.
  • serial_dev — reports that cheaper models handle many tasks across large codebases, but sometimes spin without solving the problem; their fallback is to retry with a more capable model.
  • jayhack — offers a workplace counterexample outside coding: for nontechnical knowledge work, they say other models are at least as competitive on the measures their employer cares about.

HN handles are pseudonymous; HN provides no per-comment scores. The order is HN’s ranking, and this is a slice of the thread, not a consensus.