Pyyan / Compare / Kimi K3 vs DeepSeek V4-Pro vs GLM-5.2 vs Llama 5
Kimi K3 vs DeepSeek V4-Pro vs GLM-5.2 vs Llama 5
4 of 5Open-Weight Models · verified 13 Aug 2026
| Specification | Kimi K3 | DeepSeek V4-Pro | GLM-5.2 | Llama 5 |
|---|---|---|---|---|
| Summary | Largest open-weight model shipped. | Best all-round open model of 2026. | Strongest open reasoning and mathematics. | Open-weight flagship with recursive self-improvement. |
| Licence | Open weights | MIT | Open weights | Llama licence |
| Parameters | 2.8T | — | — | 600B+ |
| Released | 17 Jul 2026 | — | — | — |
| Category | Open-Weight Models | Open-Weight Models | Open-Weight Models | Open-Weight Models |
| Official | Moonshot AI ↗ | DeepSeek ↗ | Zhipu AI ↗ | Meta ↗ |
Highlighted rows are where these differ.
DeepSeek V4-Pro
- Tops open leaderboards on agentic coding and reasoning
- MIT licence — no conditions
- Pioneered aggressive cache-hit pricing at ~$0.07/M
Best for everything open.
Llama 5
- Refines its own reasoning and generates synthetic training data
Best for open frontier.