BenchLeader

DeepSeek V3 0324

DeepSeek V3 0324 is a DeepSeek open-weights model, released 24 Mar 2025. Its best configuration ranks #284 of 777 on the BenchLeader Index at 53.5 ±0.9, in the upper half. The ± is the point: 275 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in human preference (57) and lowest in maths (54). Last measured 20 Oct 2026.

Blended price
–
Output speed
–
First answer
–
Context
–
Full answer
–
Cost per run
–
Released
24 Mar 2025
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index54
  2. Reasoning56
  3. Coding55
  4. Maths54
  5. Human preference57

Versions

DeepSeek has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.

ModelReleasedIndexRank
DeepSeek V3 0324this page24 Mar 202553.5#284
DeepSeek V326 Dec 202444.0#549

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmarknot statedSource
Arena Hard Prompts1408#164Arena (LMArena)

Coding

Maths

Benchmarknot statedSource
Arena Maths1367#183Arena (LMArena)

Instruction following

Benchmarknot statedSource
Arena Instruction Followingnot in index1378#164Arena (LMArena)

Human preference

Benchmarknot statedSource
Arena Text1395#153Arena (LMArena)
Arena Creative Writingnot in index1390#121Arena (LMArena)
Arena Multi-turnnot in index1411#139Arena (LMArena)
Arena Longer Queriesnot in index1394#165Arena (LMArena)

More from DeepSeek

New this month

See also

Data as of 21 Oct 2026. Compare with another model.

Cite as: BenchLeader, “DeepSeek V3 0324: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/deepseek-v3-0324, data as of 21 Oct 2026.