DeepSeek V4 Pro vs Kimi K3.1 Pro
Compare models
DeepSeek V4 Pro is more worth it — 62% vs 31%
How we score worth-it? →Latest DeepSeek — V4.1 with a Flash variant
- Deepseek blazes through work - at quite good quality - and you don't find yourself waiting for your agents to finish... Being able to see the CoT and not being randomly blocked for cyber review.— hn
- Feels better to me in the sense that it doesn't try to be too clever. When I ask it to do things it actually generates code that a developer would write. That still lets me build software in layers like I do mentally. And it keeps you in control of where abstractions are.— hn
- DS flash has been so consistently good I've kinda settled on it for 90% of my work... I genuinely feel as tho if no new models ever get made I'd be fine with just using DS forever.— hn
- I second this, I love DS and am not looking back. For Elixir, not a lot of models are good, but DS just blazes through it.— hn
- I ported some of my childhood Finnish shareware games for Linux and Rust with DeepSeek V4.1 Flash. It took about six hours of spelunking and finally I was pretty happy with the results! I did it twice. Tokens for both of the games were around 3 euros.— hn
- DeepSeek v4.1 Flash hallucinates like crazy.— hn
- Deepseek-v4-pro is 1.6T params, the cut-down version is 284B params, and I wouldn't trust them to do something on their own.— hn
- There's no reason one would choose Deepseek V4 Pro 0813 (max) even though it's in the 'most attractive quadrant', because GLM-5.3-Flash is both cheaper and scores better.— hn
- My biggest annoyance is that DeepSeek does seem to be considerably rate-limited of late — the agent takes 10-20 minutes to do stuff that takes much less time via OpenRouter.— hn
- Same project, same prompt — DeepSeek V4 Pro produced blank output. DS wasted 29.49 yuan in total — a real waste. Without turning on Opus it really can't compete.— v2ex ↗
“Flash hits 300 tokens/sec on real tasks — feels instant”
“Heavy agent use ran only ¥20/day — great value”
“Some devs hit bad lag on the hosted API and gave up on it”
2.8T params, beats GPT-6 Astra on Terminal-Bench
- It does pretty well for the price, around Sonnet 5.5.— hn ↗
- On mobile Kimi definitely wins from the UX point of view— hn ↗
- Its strengths are coding and multi-step agent tasks (Terminal-Bench 4.0 up from 13% to ~67%).— v2ex ↗
- Artificial Analysis Intelligence Index 55, ranked 4th globally, 11 points higher than K3; Terminal-Bench 4.0 about 67%, the highest of any model.— v2ex ↗
- Genuinely honest about what it didn't build, security-conscious, and handled the keyboard protocol better than most models tested.— youtube ↗
- I bought a yearly subscription to kimi when it was the hype and never use it. The 5h limit makes it completely unusable. You barely start doing anything and you hit the limit immediately. I advise everyone against using kimi.— hn ↗
- Kimi didn't even finish the scan, ran out of the 5hr budget in less than 10 minutes. Unusable, unfortunately.— hn ↗
- The subscription side has more complaints — people say the plan runs out within tens of minutes.— v2ex ↗
- For writing and roleplay, quite a few people report it's worse than K3.— v2ex ↗
- If you've got money to burn, go ahead and use it.— v2ex
“The 1M context actually works. I dumped a whole codebase in and it kept track of everything”
“Terminal-Bench scores beating the top frontier models is not a fluke — it's genuinely that good at agent work”
“Half the price of the top frontier tier for comparable quality. The value is unreal”
FAQ
Is DeepSeek V4 Pro or Kimi K3.1 Pro more worth it?
DeepSeek V4 Pro is more worth it, with a 62% worth-it score vs 31% for Kimi K3.1 Pro.
How much does DeepSeek V4 Pro cost vs Kimi K3.1 Pro?
DeepSeek V4 Pro: $0.66/$1.98 per 1M (off-peak) · Kimi K3.1 Pro: $2/$10 per 1M.
What do users say about DeepSeek V4 Pro vs Kimi K3.1 Pro?
DeepSeek V4 Pro: Flash hits 300 tokens/sec on real tasks — feels instant. Kimi K3.1 Pro: The 1M context actually works. I dumped a whole codebase in and it kept track of everything.