Back to rankings

Model comparison

Selected 2 / 5

See the differences. Choose what matters to you.

Differences vs. k2.8 · high

Compare settings

Comparing tested reasoning efforts of the same model

Score, cost & time differences

Current baseline · k2.8 · high

70Knowledge & Reasoning

Compared with k2.8 · high

  • k2.8 · max

    Knowledge & Reasoning: tiedReference cost $0.59 ($0.39 more)Evaluation time 53m 24s (28m 40s more)

Differences reflect these evaluation tasks. Cost and time cover the same dimension, not a typical conversation. Scoring & cost notes

Tested configuration
Effort
Effort
Metrick2.8 · highk2.8 · max
Knowledge & Reasoningk2.8high
70
k2.8max
70Tied
Overallk2.8high
63.55
k2.8max
—
Backend engineeringk2.8high
54
k2.8max
—
Frontend interactionk2.8high
72
k2.8max
—
Visual aestheticsk2.8high
57
k2.8max
—
Maintainabilityk2.8high
53
k2.8max
—

Next, validate with your own work

Try a familiar task with each candidate. Compare the results, revision effort, and actual consumption.

Explore subscriptions for these modelsUsing an API? Refer to the evaluation consumption above. For subscriptions, explore matching plans and estimated runtime.