DeepSeek V4 Pro
DeepSeek V4 Pro
A reasoning model at the frontier, published under MIT, priced like an afterthought.
Updated
The read
Pro is DeepSeek's reasoning flagship: a very large sparse model that thinks before it answers and lands close to the best closed models on mathematics and code. The remarkable part is not the benchmark, it is that the weights are MIT-licensed and the hosted API still costs a fraction of any Western equivalent. It is slower than Flash, and it should be, since the thinking is the product.
Where it shines
- Frontier-adjacent reasoning at open-weights prices
- MIT licence, including for the reasoning traces
- Very long context, so whole repositories fit in one call
Worth knowing first
- Thinking time makes it the wrong choice for interactive work
- Self-hosting the full model is a serious hardware commitment
Index score
86.8
Provisional. This score is read from the product's public capability, not from the one-prompt rebuild the fully reviewed entries went through, so treat it as a placing rather than a verdict.
- Rank
- 07 of 53
- Builds
- LLMs
- Category
- Reasoning
- Output
- MIT-licensed weights, plus a cheap API
- Pricing
- Free tier
- Best for
- Hard reasoning without a frontier-lab invoice
Scores are our own editorial judgement, weighted craft 55, speed 10, control 15 and value 20.
Compare it with
Also worth a look
Full index- 02GPT-5.6 SolGPT-5.6 SolProvisionalOpenAI's deepest thinker: the one to hand a problem you cannot solve yourself.Reasoning modelPaid only88.6
- 26Hunyuan HY3Hunyuan HY3ProvisionalSwitches between fast answers and slow thinking, and charges almost nothing for either.Reasoning modelPaid only80.0
- 47Perplexity Sonar ProSonar ProProvisionalA model with the live web wired in, and every claim comes back with a source attached.Reasoning modelPaid only71.6
