DeepSeek-R1 at a glance
DeepSeek-R1 claimed reasoning performance on par with OpenAI o1, shipped with MIT-licensed weights, six distilled smaller models and a low-cost API.
What changed with DeepSeek-R1
On January 20, 2025, DeepSeek released DeepSeek-R1, a reasoning model it said performs on par with OpenAI o1 on math, code and reasoning tasks. The model, its weights and a technical report were published openly, and the company relicensed R1 under the MIT license so that weights and outputs can be used for fine-tuning and distillation.
DeepSeek credited large-scale reinforcement learning in post-training for most of the gains. It also open-sourced six smaller models distilled from R1, saying the 32B and 70B versions are comparable to OpenAI o1-mini. In the API the model is called deepseek-reasoner, priced at $0.55 per million input tokens on a cache miss ($0.14 on a cache hit) and $2.19 per million output tokens.
DeepSeek-R1 benchmarks and pricing
Values are copied from the linked source on the date shown. They are not IA Compa measurements.
- Input price per 1M tokensas of · Source for Input price per 1M tokens: api-docs.deepseek.com (opens in a new tab)
- $0.55 (cache miss)
- Output price per 1M tokensas of · Source for Output price per 1M tokens: api-docs.deepseek.com (opens in a new tab)
- $2.19
Where to compare DeepSeek-R1
Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.
General leaderboards
Why DeepSeek-R1 matters for businesses
For businesses, R1 showed that frontier-style reasoning no longer required a closed model or premium pricing, which pushed the whole market toward cheaper reasoning. Teams can self-host the weights to keep data in-house, but should still check licensing of distilled bases and data-residency rules before using the hosted API.
Sources for this story
- DeepSeek-R1 Release (opens in a new tab)Official
DeepSeek API Docs · api-docs.deepseek.com
- DeepSeek-R1 repository (opens in a new tab)Official
GitHub · github.com
- DeepSeek R1 (Jan) model page (opens in a new tab)Benchmark
Artificial Analysis · artificialanalysis.ai
Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.
- #llm
- #deepseek
- #open-weights
- #reasoning