top 10 modes - chunhualiao/public-docs GitHub Wiki
OpenRouter Pricing Comparison (as of February 2026)
| Rank (Original) | Model | Provider | Input Price ($/M tokens) | Output Price ($/M tokens) | Context Length | Notes / Key Features |
|---|---|---|---|---|---|---|
| 1 | Gemini 3 Flash Preview | $0.50 | $3.00 | 1.05M | High-speed thinking model, agentic workflows, multimodal (text/images/audio/video/PDF) | |
| 2 | Claude Sonnet 4.5 | Anthropic | $3.00 | $15.00 | 1M (beta) / 200K (GA) | Hybrid reasoning for agents & coding; 1M available with beta header |
| 6 | Gemini 2.5 Flash | $0.30 | $2.50 | 1M+ | Workhorse for reasoning, coding, math; built-in thinking | |
| 9 | Grok 4.1 Fast | xAI | $0.20 | $0.50 | 2M | Best for agentic tool calling, real-world tasks like support/research |
| 10 | Gemini 2.5 Flash Lite | $0.10 – $0.13 | $0.40 – $0.50 | 1M+ | Ultra-low cost & latency variant, high-volume / long-context use |
Filtered Out (Context < 1M)
- DeepSeek V3.2 (~128K–164K)
- Kimi K2.5 (~256K–262K)
- Claude Opus 4.5 (~200K)
- MiniMax M2.1 (~197K or less)
- Grok Code Fast 1 (likely <1M, specialized coding focus)
Prices are per 1 million tokens and based on the most recent available data from OpenRouter model pages and provider docs.
Note: Prices may vary slightly by provider routing, context length thresholds (>128K or >200K tokens often double), or caching. These are standard rates (no caching discounts applied unless noted). All models are available via OpenRouter.
| Rank | Model | Provider | Input Price ($/M tokens) | Output Price ($/M tokens) | Context Length | Notes / Key Features |
|---|---|---|---|---|---|---|
| 1 | Gemini 3 Flash Preview | $0.50 | $3.00 | 1M+ | High-speed thinking model, agentic workflows | |
| 2 | Claude Sonnet 4.5 | Anthropic | $3.00 | $15.00 | 1M | Optimized for agents & coding |
| 3 | DeepSeek V3.2 | DeepSeek | $0.25 – $0.28 | $0.38 – $0.42 | ~164K | Extremely cost-effective reasoning |
| 4 | Kimi K2.5 | Moonshot AI | $0.45 – $0.60 | $2.50 – $3.00 | 262K | Multimodal, strong coding |
| 5 | Claude Opus 4.5 | Anthropic | $5.00 | $25.00 | 200K | Frontier reasoning model |
| 6 | Gemini 2.5 Flash | $0.30 | $2.50 | 1M+ | Workhorse for reasoning & coding | |
| 7 | MiniMax M2.1 | MiniMax | $0.27 | $0.95 – $1.20 | ~197K | Efficient coding & agentic workflows |
| 8 | Grok Code Fast 1 | xAI | $0.20 | $1.50 | 256K | Specialized for coding tasks |
| 9 | Grok 4.1 Fast | xAI | $0.20 | $0.50 | 2M | Best for agentic tool calling |
| 10 | Gemini 2.5 Flash Lite | $0.10 – $0.13 | $0.40 – $0.50 | 1M+ | Ultra-low cost & latency variant |
Key Insights
- Cheapest overall: Gemini 2.5 Flash Lite and Grok 4.1 Fast (input ~$0.10–$0.20, output ~$0.40–$0.50) — great for high-volume or long-context use.
- Best value for performance: DeepSeek V3.2 and MiniMax M2.1 offer strong reasoning/coding at under $1/M output.
- Premium frontier models: Claude Opus 4.5 and Sonnet 4.5 are more expensive but excel in complex agentic tasks.
- Google models dominate low-cost high-context options (e.g., Gemini series with 1M+ context).
- xAI models (Grok) provide excellent tool-calling and low pricing for agentic workflows.
Prices can fluctuate slightly due to OpenRouter routing to different providers. For the latest exact rates, check the individual model pages on OpenRouter.ai/models.