Kimi K3 vs GLM-5.3
Kimi K3 is from Moonshot AI, released July 16, 2026. GLM-5.3 is from Zhipu AI, released August 18, 2026.
- LMArena Elo1488 vs 1481. Kimi K3 is 0.5% ahead on LMArena Elo.
- Input per 1M$3 vs $1.40. GLM-5.3 costs 53% less per million input tokens.
- Output per 1M$15 vs $4.40. GLM-5.3 costs 71% less per million output tokens.
- Context window1M vs 1M. Both hold 1M tokens of context.
Prices are the vendors’ published API rates; Elo is LMArena’s text board. Each figure is dated on its model page.
Interactive Comparison
Loading chart...
$3 / $15 per million tokens | $1.40 / $4.40 per million tokens | |
|---|---|---|
| Identity | ||
| Developer | Moonshot AI | Zhipu AI |
| Released | Jul 2026 | Aug 2026 |
| Status | GA | GA |
| Licence | Open weights (Kimi K3 License) | GLM-5.3 License (custom: MIT-style terms plus a security-review clause for MaaS providers above $10B revenue); weights on Hugging Face, zai-org/GLM-5.3 |
| Self-hostable | Yes | Yes - weights published on Hugging Face |
| Cost | ||
| Blended $/1M tokensinput × 0.75 + output × 0.25 | $6.000 / 1M tokens | $2.150 / 1M tokens |
| Input price | $3.000 / 1M tokens | $1.400 / 1M tokens |
| Output price | $15.000 / 1M tokens | $4.400 / 1M tokens |
| Cached input | $0.30 / 1M tokens (automatic context-cache hit; $3.00 on a cache miss) | $0.26 / 1M tokens |
| Batch discount | Not offered | Not offered; GLM Coding Plan applies a 50% off-peak points reduction |
| Free tier | Not offered | Not offered |
| Capacity | ||
| Context window | 1M tokens | 1M tokens |
| Max output | 128k tokens (default; max_completion_tokens configurable up to 1M) | 128k tokens |
| Long-context surcharge | None (flat rate across the 1M window) | Not published |
| Capability | ||
| Vision in | Yes | No — text-only input |
| Audio in | No | No |
| Function calling | Yes | Yes |
| Structured output | Yes | Yes |
| Extended reasoning | Yes | Yes — always on, three effort levels (low, high, max) |
| Web search | Yes (tool currently being updated by Moonshot; not recommended for near-term use) | Not documented |
| Code execution | Not offered | Not documented |
| Access | ||
| API | Yes | Yes |
| Chat app | Yes | Yes |
| Cloud marketplaces | Azure AI Foundry (via Fireworks AI), OpenRouter | Not offered |
| Fine-tuning | Not offered | Not offered |
| Measured quality | ||
| LMArena Elo | 1488 (checked Nov 2026) | 1481 (checked Nov 2026) |
| LMArena rank | Rank 19 | Rank 28 |
| GPQA Diamond | 93 | 91 |
| Elo per dollarLMArena Elo ÷ blended $/1M tokens | 248 | 689 |
