Atria Dawn Preview vs SWE-2: Specs & Benchmark Comparison

Atria Dawn Preview is developed by Shanghai AI Laboratory, while SWE-2 comes from Cognition. Both were released in September 2026. Atria Dawn Preview has a published size of about 744 billion parameters; Cognition has not disclosed the parameter count of SWE-2.

We have 12 benchmark results for Atria Dawn Preview and 4 benchmark results for SWE-2, but they overlap on a single test: Terminal-Bench 2.1, where SWE-2 scores 92.8% against 78.3%.

CharacteristicAtria Dawn PreviewSWE-2
CompanyShanghai AI LaboratoryCognition
Release DateSeptember 11, 2026September 10, 2026
Parameters744B—
MultimodalNoNo
Average Score72.7%60.8%
Benchmarks
Terminal-Bench 2.178.3%92.8%

Visual Benchmark Comparison

Atria Dawn Preview
SWE-2
Terminal-Bench 2.10.8 vs 0.9
0.8
0.9

Verdict

Both models show equal results — the choice depends on your specific use case.

Overall Performance

Both models show comparable average scores: Atria Dawn Preview — 0.7, SWE-2 — 0.6.

Recency

Both models were released around the same time: 9/11/2026 and 9/10/2026.

More About These Models

Related Comparisons

Frequently Asked Questions

Which is better for coding — Atria Dawn Preview or SWE-2?
Direct comparison on the SWE-Bench benchmark is not available. We recommend reviewing other metrics on the comparison page.
Which model is cheaper — Atria Dawn Preview or SWE-2?
API pricing data is available on the individual model pages.
Which has a larger context window — Atria Dawn Preview or SWE-2?
Context window data is available on the individual model pages.

The Atria Dawn Preview and SWE-2 comparison is updated for 2026. Data includes benchmark results, API pricing, context window size and other specifications. For more detailed information, visit the Atria Dawn Preview or SWE-2 page. See also the complete list of AI model comparisons.