BenchLeader

DeepSeek-R1-Distill-Qwen-14B

Best configuration ranks #485 of 610 on the BenchLeader Index at 41.4 ±9.9. Last measured 20 Jan 2025stale: no new result in six months. Released 20 Jan 2025.

Blended price
$0.220/M
$0.150 in · $0.431 out
Output speed
First answer
Context
33k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index41
  2. Reasoning35
  3. Maths55
  4. Instruction following27
  5. Long context30
  6. Composite39

Benchmark results

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

Instruction following

BenchmarkdefaultSource
IFBench22.1%#384Artificial Analysis

Long context

BenchmarkdefaultSource
AA-LCR10.3%#394Artificial Analysis

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0002
Summarise a 30-page report12,000 / 600$0.0021
Code edit6,000 / 1,500$0.0015
Agentic coding session60,000 / 4,000$0.011
Structured extraction2,000 / 200$0.0004

See also

Data as of 9 Sept 2026. Compare with another model.