BenchLeader
Deep CogitoOpen weightsReasoning modelAuto-detected

Cogito v2.1

Not yet ranked: too few independent results so far.

Blended price
$1.25/M
$1.25 in · $1.25 out
Output speed
First answer
Context
128k
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Agents & tools48
  2. Knowledge52
  3. Instruction following48
  4. Long context36

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

BenchmarkthinkingSource
GPQA Diamond (AA)not in index76.8%#220Artificial Analysis
Humanity's Last Exam (AA)not in index12.0%#221Artificial Analysis

Agents & tools

BenchmarkthinkingSource
Terminal-Bench Hard16.7%#179Artificial Analysis

Knowledge

BenchmarkthinkingSource
AA-Omniscience-25.8#200Artificial Analysis

Instruction following

BenchmarkthinkingSource
IFBench46.3%#191Artificial Analysis

Long context

BenchmarkthinkingSource
AA-LCR22.7%#348Artificial 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$0.0009
Summarise a 30-page report12,000 / 600$0.016
Code edit6,000 / 1,500$0.0094
Agentic coding session60,000 / 4,000$0.080
Structured extraction2,000 / 200$0.0028

See also

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