DeepSeek Reasoner

DeepSeek · DeepSeek V3.2 · budget · Current model version: DeepSeek-V3.2 thinking mode · deprecated Jul 24, 2026 · deepseek-reasoner

DeepSeek V3.2 family:DeepSeek Chat·DeepSeek Reasoner

DeepSeek's thinking lane with the same low price structure and a much larger output ceiling than chat mode.

Pricing

Input$0.028 / 1M cache hit, $0.28 / 1M cache miss
Cached input$0.028 / 1M
Output$0.42 / 1M

Same published rates as chat mode on the referenced pricing page.

Cost bandUltra-budget
Free tierPaid API.

Limits

Context window128k
Max output64k
Knowledge cutoffNot surfaced in pricing page
DeploymentDeepSeek API
OpenAI-compatible APIYes
ModalitiesText

Where it fits

Best forCheap reasoning runs, structured analysis, and experiments where you want maximal tokens per dollar.
Business fitCompelling on price, but selection should include governance and platform-risk review.

Capabilities

JSON outputTool callsPrefix completion

Benchmark scores

Composite88 / 100
Coding84 / 100 (SWE-bench Verified)
Reasoning96 / 100 (GPQA Diamond)
Long context80 / 100 (RULER 64K)
Vision64 / 100
Instruction following82 / 100 (IFEval)

Editorial aggregate of published benchmark results as of Oct 1, 2025. These are not independent measurements run by this site.

Change history

Data

Last checkedMar 31, 2026

Verify with provider documentation before committing spend. Report an error and it gets fixed with a logged correction.

Open in the comparison table JSON API