Compare
DeepSeek-V4-Pro-0813 vs KAT-Coder-Air V2.5
DeepSeek-V4-Pro-0813 (DeepSeek) and KAT-Coder-Air V2.5 (Kwaipilot) compared on benchmarks, pricing, context window, and use-case rankings.
- KAT-Coder-Air V2.5 is cheaper on input tokens — $0.15 vs $0.58 per 1M.
- DeepSeek-V4-Pro-0813 has a larger context window (1M vs 256K tokens).
| Attribute | DeepSeek-V4-Pro-0813 | KAT-Coder-Air V2.5 |
|---|---|---|
| Scores | ||
| Intelligence (ECI) | 155 | — |
| Coding | 78 | — |
| Math | 88 | — |
| Reasoning & Knowledge | 70 | — |
| Agentic & Tools | — | — |
| Specifications | ||
| Developer | DeepSeek | Kwaipilot |
| Family | DeepSeek | — |
| Released | Aug 13, 2026 | Jul 10, 2026 |
| Parameters | 1.6T | — |
| Availability | Open weights (unrestricted) | — |
| Context window | 1M | 256K |
| Price — $/M input | $0.58 | $0.15 |
| Price — $/M output | $1.74 | $0.60 |
| Inputs | text | text |
| Outputs | text | text |
| Benchmarks | ||
| AIME 2024/2025 | 99% | — |
| ARC-AGI | 91% | — |
| ARC-AGI-2 | 61% | — |
| GPQA Diamond | 92% | — |
| Humanity's Last Exam | 41% | — |
| SciCode | 51% | — |
| SimpleQA Verified | 53% | — |
| WebDev Arena | 1582 | — |
| WeirdML | 66% | — |
Use-case scores are 0–100 percentile composites across each area’s benchmarks, ranked against every model from the past year. Highlighted cells lead each row. Open a model for the full picture.
Frequently asked questions
Which is cheaper, DeepSeek-V4-Pro-0813 or KAT-Coder-Air V2.5?
KAT-Coder-Air V2.5 is cheaper on input tokens at $0.15 per million, versus $0.58 (representative OpenRouter pricing).
Which has a larger context window, DeepSeek-V4-Pro-0813 or KAT-Coder-Air V2.5?
DeepSeek-V4-Pro-0813 supports up to 1M tokens, compared with 256K for the other.
Want a different match-up? Open the compare tool to add or swap models.
More comparisons
Benchmarks & model data from Epoch AI (CC BY); pricing & specs from OpenRouter. ECI = Epoch Capabilities Index.