BenchLeader
Multiverse ComputingOpen weightsReasoning modelAuto-detected

HyperNova 60B 2605 (high, based on gpt-oss-120b)

Best configuration ranks #308 of 610 on the BenchLeader Index at 49.4 ±8.4. Released 26 May 2026.

Blended price
$0.065/M
$0.040 in · $0.140 out
Output speed
345 tok/s
First answer
6.48 s
Context
131k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index49
  2. Agents & tools54
  3. Knowledge38
  4. Instruction following65
  5. Long context44
  6. Composite44

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 index73.3%#258Artificial Analysis
Humanity's Last Exam (AA)not in index16.6%#182Artificial Analysis

Agents & tools

Knowledge

BenchmarkdefaultSource
AA-Omniscience-54.6#357Artificial Analysis

Instruction following

BenchmarkdefaultSource
IFBench66.5%#95Artificial Analysis

Long context

BenchmarkdefaultSource
AA-LCR36.3%#301Artificial Analysis

Composite

BenchmarkdefaultSource
AA Intelligence Index11.7#279Artificial 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$0.00017.3 s
Summarise a 30-page report12,000 / 600$0.00068.2 s
Code edit6,000 / 1,500$0.000510.8 s
Agentic coding session60,000 / 4,000$0.003018.1 s
Structured extraction2,000 / 200$0.00017.1 s

See also

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