Cognition's SWE-2

Cognition's coding model, 64% cheaper than Fable 5.1

View product page →

Editorial score

26.6

93th percentile of this pass

Evidence coverage

20%

6 of 8 dimensions measured

Model confidence

70%

openrouter

Surprise

attention vs fundamentals

Platform divergence

+30

no meaningful disagreement

Investigate

Ranks high on little evidence. Worth a look, but verify before committing.

How the score was built

Unmeasured dimensions are excluded from the average rather than counted as zero, so the remaining weights are renormalised. The share column is the weight each one actually carried.

Dimension	Value	Base weight	Actual share
Trend	50	20%	25%
Story	80	20%	25%
Novelty	89	15%	19%
Business	0	15%	19%
Revenue	not measured	10%	—
Founder	not measured	10%	—
Cross-platform	86	5%	6%
Timeliness	100	5%	6%

Not measured: Revenue, Founder. These are gaps in our data, not findings about the product.

Why this is a story

InterestingNot necessarily successful — exceptional idea or story.

SurprisingThe idea itself is unusual or unexpected. · primary

Evidence

Cognition's SWE-2 is described as 64% cheaper than Fable 5.1.

Model brief

Cognition's SWE-2 model is 64% cheaper than Fable 5.1.

What happened: Cognition launched its SWE-2 coding model, positioning it as a cheaper alternative to Fable 5.1.
Why it's interesting: The pricing strategy of 64% cheaper than a competitor is a notable departure in this market.
The bigger story: This launch highlights a trend of aggressive pricing in AI coding models, which could reshape competition.
Why now: Cognition just launched SWE-2, drawing significant attention on Hacker News.

scored 2026-09-13T08:39:00.287Z · classified 2026-09-13T08:39:00.287Z