ColinBuilds.com · Open-weight R1 profile · provider pricing only
DeepSeek-R1-0528
- Provider-listed context window
- 163,840 tokens
Official pricing status
No current official DeepSeek API price is recorded for an R1 API route. DeepSeek's current API docs list deepseek-v4-flash and deepseek-v4-pro, while deepseek-reasoner is deprecated and currently routes to deepseek-v4-flash thinking mode. OpenRouter pricing is provider-hosted, not official DeepSeek creator pricing.
Compare models
Same per-1M input/output rates as above. Enter your monthly token volumes to compare bills.
DeepSeek-R1-0528
- Input: 1M × $0.50 = $0.50
- Output: 1M × $2.15 = $2.15
- Total
- $2.65
GPT-4o
- Input: 1M × $2.50 = $2.50
- Output: 1M × $10.00 = $10.00
- Total
- $12.50
Difference
$9.85
GPT-4o costs $9.85 more than DeepSeek-R1-0528.
Benchmark Matrix
Reported model metrics from published sources.
Active Parameters
37B
Reported active parameter count at inference time.
Popular comparisons
30 total featuring DeepSeek-R1-0528.
DeepSeek-R1-0528 is the current DeepSeek R1 open-weight update checked here. It should not be confused with the current official DeepSeek API model IDs, which are now V4 routes.
What this model is
DeepSeek’s R1 line is its open-weight reasoning model family. The DeepSeek-R1-0528 model card says the R1 model received a minor version upgrade, with improved reasoning depth, reduced hallucination rate, function-calling support, and better coding experience.
For API use, the current DeepSeek docs need careful wording. The live Models & Pricing page lists deepseek-v4-flash and deepseek-v4-pro as the currently supported model IDs. The same pricing page says deepseek-chat and deepseek-reasoner will be deprecated on 2026-07-24 15:59 UTC and currently correspond to deepseek-v4-flash non-thinking and thinking modes. That means this page does not present deepseek-reasoner as a current official R1 API price.
Pricing notes
No current official DeepSeek API price is recorded here for a DeepSeek-R1 API route. The historical DeepSeek-R1 release post used model=deepseek-reasoner, but the live API docs now route that compatibility name to V4-Flash thinking mode and mark it for retirement.
The calculator on this page uses provider-hosted OpenRouter pricing for deepseek/deepseek-r1-0528: $0.50 per 1M input tokens and $2.15 per 1M output tokens, checked on 2026-07-07 through OpenRouter’s public models API. OpenRouter also lists an input cache-read price of $0.35 per 1M tokens, but the calculator uses the non-cache input rate for safer comparison.
OpenRouter pricing is provider-hosted marketplace pricing for deepseek/deepseek-r1-0528; it is not official DeepSeek creator/API pricing. Hosted R1 costs can vary by provider and route.
Benchmarks and specs
The DeepSeek-R1-0528 model card says the current R1 version improves AIME 2025 accuracy from 70% in the previous version to 87.5%, and that average reasoning length on AIME increased from 12K tokens to 23K tokens.
The 37B active-parameter value is sourced from DeepSeek’s original R1 architecture README and provider metadata, not from a current official DeepSeek API R1 route. The 163,840-token context is OpenRouter/provider-listed; DeepSeek’s original R1 README listed 128K context for the original R1.
Best fit
DeepSeek-R1-0528 is best for reasoning-focused comparisons where open weights matter and where the price shown is clearly labelled as provider-hosted.
It is not the cleanest choice for a current official DeepSeek API pricing comparison. For that, DeepSeek’s live docs point to the V4 model IDs, especially V4-Pro and V4-Flash with thinking mode controls.