BenchLeader
Swiss AI InitiativeOpen weightsAuto-detected

Apertus 70B Instruct

Best configuration ranks #546 of 610 on the BenchLeader Index at 38.6 ±3.5.

Blended price
$1.34/M
$0.820 in · $2.92 out
Output speed
First answer
Context
66k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index39
  2. Agents & tools34
  3. Knowledge35
  4. Instruction following30
  5. Long context25
  6. Composite35

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

BenchmarkdefaultSource
GPQA Diamond (AA)not in index27.2%#503Artificial Analysis
Humanity's Last Exam (AA)not in index5.5%#349Artificial Analysis

Agents & tools

Knowledge

BenchmarkdefaultSource
AA-Omniscience-61.5#399Artificial Analysis

Instruction following

BenchmarkdefaultSource
IFBench25.9%#371Artificial Analysis

Long context

BenchmarkdefaultSource
AA-LCR0.0%#419Artificial Analysis

Composite

BenchmarkdefaultSource
AA Intelligence Index5.1#514Artificial 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$0.0012
Summarise a 30-page report12,000 / 600$0.012
Code edit6,000 / 1,500$0.0093
Agentic coding session60,000 / 4,000$0.061
Structured extraction2,000 / 200$0.0022

See also

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