DeepSeek V4 Pro
DeepSeek's larger V4-series large language model, available through its pay-per-token API.
HN commenters use it as a very cheap daily driver for coding and agent work,12345 with costs kept low by deep discounts on cache reads.6789 The most common caveat is that it makes mistakes more often than top models,101112 and some users switch to a pricier model when it gets stuck.1314
Why people recommend it
Low cost · 3
Commenters repeatedly call it cheap enough to use for nearly everything, including trivial tasks.
“It's also unbelievably cheap - so cheap, in fact, that I have started using it even for trivial tasks.”
“For hobby projects I've completely switched to DeepSeek v4 pro.”
Show 1 more
“are genuinely good enough as daily drivers, given the costs”
Cache pricing · 3
Low cache-read prices make long agentic sessions and big-context code analysis very cheap.
“it's actually ~60x cheaper if you take into account the typical cache-read/input/output split”
“I'm guessing this is almost entirely about the incredibly low cache read prices.”
Show 1 more
“very cheap large-scale code analyses and things that require mega context windows”
Daily workhorse · 3
Several users run it almost exclusively as their main model for coding and general work.
“I have been using deepseek v4 pro almost exclusively.”
“Yeah deepseek is my workhorse of choice these days”
Show 1 more
“I main V4 Pro at work now, and at home I route between Pro and Flash based on task.”
Pairs with Flash · 3
Users route harder planning work to Pro and cheaper detail work to V4 Flash.
“I main V4 Pro at work now, and at home I route between Pro and Flash based on task.”
Show 1 more
“an occasional usage of DeepSeek V4 Pro”
Follows instructions · 3
Some use it to implement specs written elsewhere because it does what it is told.
“have Deepseek v4 Pro implement because it just seems to actually do what it's told, given enough tries”
“Deepseek is far and away the most cost effective, fast, and quite good for 90% of implementation work”
Show 1 more
“makes DeepSeek Pro the best LLM for me. It allows me to work relatively fast at a very low price.”
Writing quality · 1
One commenter finds its writing more nuanced and natural.
Caveats and dissent
Frequent mistakes · 2
Some users report regular errors that took extra time to fix.
“the preview version of the model kept regularly messing things up”
“DeepSeek V4 Pro is really affordable per-token but regularly kept making mistakes in the tasks I gave it.”
Fallback needed · 2
When it gets stuck, users switch to GPT models to find the problem.
“Yeah deepseek is my workhorse of choice these days”
“I have been using deepseek v4 pro almost exclusively.”
Big tasks · 2
Newer OpenAI, Anthropic and Claude Opus models are seen as better for complex tasks.
“makes DeepSeek Pro the best LLM for me. It allows me to work relatively fast at a very low price.”
Slow · 2
One tester found Pro really slow compared with Flash.
“Yeah deepseek is my workhorse of choice these days”
API vs subscription · 2
Paying API prices can cost more than a flat subscription to another model.
“I tried for example Deepseek V4 Pro for a bit (via API pricing), but it's _more expensive_ than Sol via subscription”
“I have been using Deepseek V4 Pro for personal projects and it has been great.”
Data and drift · 2
The Reasonix harness sends data to China, and very large contexts bring some attentional drift.
“very cheap large-scale code analyses and things that require mega context windows”
“very cheap large-scale code analyses and things that require mega context windows”
Often compared with
2 moved from Claude Opus to DeepSeek V4 Pro
1 moved from DeepSeek V4 Pro to DeepSeek V4 FlashGLM41 comments mention both
1 moved from GLM to DeepSeek V4 Pro
1 moved from DeepSeek V4 Pro to GLMKimi31 comments mention both
2 moved from Kimi to DeepSeek V4 Pro
5 moved from Claude to DeepSeek V4 Pro
2 moved from Claude Code to DeepSeek V4 Pro
Who recommends it in their own words
Threads where it came up
- DeepSeek makes the V4 Pro price discount permanent 13 mentions
- DeepSeek V4 Pro 0813 11 mentions
- DeepSeek V4 Pro beats GPT-5.5 Pro on precision 9 mentions
- DeepSeek V4–almost on the frontier, a fraction of the price 9 mentions
- DeepSeek v4 9 mentions
- DeepClaude – Claude Code agent loop with DeepSeek V4 Pro, 17x cheaper 7 mentions
Also written as DeepSeek Pro, DeepSeek-V4-Pro, DeepSeek V4 Pro 0813, DeepSeek V4-Pro. Data: deepseek-v4-pro.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 44
“all their recent stuff (e.g. [0][1]) are MIT licensed”
damn, I might have to give it a shot myself
“very cheap large-scale code analyses and things that require mega context windows”
“very cheap large-scale code analyses and things that require mega context windows”
“I switched to DeepSeek V4 Flash and Pro for a while and I enjoyed that it was cheap enough to get work done.”
“My experience is that open weights models find between 1/4 to 1/2 of what Fable/Opus stack can find”
“and much cheaper than the current V4 Pro model”
“all other larger models I tested (DeepSeek V4 Pro, Qwen3.8 Max, GLM 5.3, Kimi K3) drew the pelican riding a bicycle”
“I tried for example Deepseek V4 Pro for a bit (via API pricing), but it's _more expensive_ than Sol via subscription”
“I main V4 Pro at work now, and at home I route between Pro and Flash based on task.”
“Deepseek v4 pro uses 2.4GB for the same context length”
“Briefly Pro is $2/1m output in off-peak periods, $4 in peak.”
“DeepSeek v4 Pro 0813 is less accurate with a 35% false positive rate versus GPT-5.6-Sol's 15% false positive rate.”
“Pro feels really slow compared.”
“For tasks like pondering on something, reviewing code, etc. I use Pro”
“Why would I use pro over flash now?”
“Seems a bit weaker than Opus 4.8 but at least 10x cheaper once you account for verbosity.”
“cheaper to use ds-v4-pro than the luna models from openai”
“it's actually ~60x cheaper if you take into account the typical cache-read/input/output split”
“pro plans, flash implements.”
“it still would be a better value to use DS”
DeepSeek V4 Flash is the "too cheap to meter" of AI. And you can run the full unquantized model locally for $8000 (2x DGX Spark) at full 1M context and decent speeds: https://github.com/elsung/dgx-spark-deepseek-v4-flash#-long-context--what-to-expect
“just now available on OpenRouter”
“Competitive with opus 4.8 but weaker than sol or fable”
“the preview version of the model kept regularly messing things up”
“I have been using deepseek v4 pro almost exclusively.”
“I'm guessing this is almost entirely about the incredibly low cache read prices.”
“High hopes for V4 Pro”
“I love deepseek's (Pro) writing.”
“makes DeepSeek Pro the best LLM for me. It allows me to work relatively fast at a very low price.”
“For DeepSeek-V4-Pro, the amount is $15.”
“cheaper per task than DS”
“here's one measurement that says v4 pro is better than luna”
“have Deepseek v4 Pro implement because it just seems to actually do what it's told, given enough tries”
“cheap Chinese models (DeepSeek Pro, Flash, MiMo 2.5 Pro for reasoning, and MiniMax M3 for agents)”
“I've found this to be true for 1M context-size open weights that I use (GLM 5.2, MiMo v2.5 Pro, & DeepSeek v4 Pro)”
> even the lower-end models can find the things they didn't think of first when inspecting the already-written output I use Deepseek V4 Pro for my hobby project (an OpenCode client). The economics will obviously be different at work. It’s worse than the GPTs, but indeed, if you focus it on reviewing its own code (essentially spending more reasoning and changing perspective), it’s also quite capable at improving its o
“Deepseek v4 Pro has three modes: no thinking, thinking ("high"), very good thinking ("max").”
“The A100 came out in 2020 and cannot run a large model like DeepSeek v4 Pro.”
“Yeah deepseek is my workhorse of choice these days”
“It's also unbelievably cheap - so cheap, in fact, that I have started using it even for trivial tasks.”
“DeepSeek V4 Pro is really affordable per-token but regularly kept making mistakes in the tasks I gave it.”
“are genuinely good enough as daily drivers, given the costs”
“For hobby projects I've completely switched to DeepSeek v4 pro.”