Vibe CodingTop-tier coding LLMs
Models
The top-tier LLMs coders are actually using right now — context windows, per-token pricing, and what each one is best at, in one table.
| Model | Provider | Context | Input / M | Output / M | Best for |
|---|---|---|---|---|---|
| Anthropic | 1M tokens | $5.00 | $25.00 | Hardest coding tasks, long autonomous agent runs | |
| Anthropic | 1M tokens | $3.00 | $15.00 | Everyday coding, best cost-to-quality ratio | |
| DeepSeek | 1M tokens | $0.44 | $0.87 | Cost-sensitive coding, self-hosting | |
| 1M tokens | $2.00 | $12.00 | Large-context, whole-codebase work | ||
| Z.ai (Zhipu) | 1M tokens | $1.40 | $4.40 | Open-weight frontier coding, self-hosting at scale | |
| OpenAI | 1.05M tokens | $5.00 | $30.00 | Frontier reasoning and coding at scale | |
| xAI | 500K tokens | $2.00 | $6.00 | Fast, tool-heavy agentic coding | |
| Moonshot AI | 256K tokens | $0.95 | $4.00 | Long-horizon coding agents, cost-sensitive frontier work | |
| MiniMax | 1M tokens | $0.60 | $2.40 | Full-repo context, multimodal coding tasks | |
| Alibaba (Qwen) | 262K tokens | $0.12 | $0.80 | Self-hosted, local coding agents |