81Fit score
This model always redirects to the latest model in the DeepSeek V4 Flash family.
Model details →Best AI models for RAG
Good RAG models have long context (to fit retrieved chunks), low hallucination, and tool use for query refinement.
81Fit score
DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows.
Model details →81Fit score
Gemini 2.0 Flash Lite offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemini Pro 1.5](/google/g…
Model details →Top 25 models for retrieval-augmented generation
| # | Model | Provider | Context | Input price / 1M | Tier |
|---|---|---|---|---|---|
| 1 | DeepSeek V4 Flash Latest | ~deepseek | 1.0M Ultra context (1M+) | $0.08 | Budget |
| 2 | DeepSeek: DeepSeek V4 Flash 0731 | DeepSeek | 1.0M Ultra context (1M+) | $0.09 | Budget |
| 3 | Google: Gemini 2.0 Flash Lite | 1.0M Ultra context (1M+) | $0.07 | Budget | |
| 4 | Google: Gemini 2.5 Flash Lite (batch) | 1.0M Ultra context (1M+) | $0.05 | Budget | |
| 5 | OpenAI: GPT-4.1 Nano (batch) | OpenAI | 1.0M Ultra context (1M+) | $0.05 | Budget |
| 6 | Auto Router | Openrouter | 2M Ultra context (1M+) | $-1000000.00 | Budget |
| 7 | Auto Router (Beta) | Openrouter | 2M Ultra context (1M+) | $-1000000.00 | Budget |
| 8 | Qwen: Qwen3.5-Flash | Qwen | 1M Ultra context (1M+) | $0.07 | Budget |
| 9 | Qwen: Qwen3.6 Plus Preview (free) | Qwen | 1M Ultra context (1M+) | $0.00 | Budget |
| 10 | Qwen: Qwen3.7 Flash | Qwen | 1M Ultra context (1M+) | $0.03 | Budget |
| 11 | Z.ai: GLM 5.2 | Z Ai | 1.0M Ultra context (1M+) | $0.10 | Budget |
| 12 | ByteDance Seed: Seed 1.6 Flash | Bytedance Seed | 262K Long context (128K+) | $0.07 | Budget |
| 13 | DeepSeek: DeepSeek V4 Flash 0423 | DeepSeek | 1.0M Ultra context (1M+) | $0.14 | Budget |
| 14 | DeepSeek: DeepSeek V4 Pro | DeepSeek | 1.0M Ultra context (1M+) | $0.43 | Budget |
| 15 | Google: Gemini 2.0 Flash | 1.0M Ultra context (1M+) | $0.10 | Budget | |
| 16 | Google: Gemini 2.5 Flash | 1.0M Ultra context (1M+) | $0.30 | Budget | |
| 17 | Google: Gemini 2.5 Flash (batch) | 1.0M Ultra context (1M+) | $0.15 | Budget | |
| 18 | Google: Gemini 2.5 Flash Lite | 1.0M Ultra context (1M+) | $0.10 | Budget | |
| 19 | Google: Gemini 2.5 Flash Lite Preview 09-2025 | 1.0M Ultra context (1M+) | $0.10 | Budget | |
| 20 | Google: Gemini 2.5 Pro (batch) | 1.0M Ultra context (1M+) | $0.63 | Budget | |
| 21 | Google: Gemini 3 Flash Preview | 1.0M Ultra context (1M+) | $0.50 | Budget | |
| 22 | Google: Gemini 3 Flash Preview (batch) | 1.0M Ultra context (1M+) | $0.25 | Budget | |
| 23 | Google: Gemini 3.1 Flash Lite | 1.0M Ultra context (1M+) | $0.25 | Budget | |
| 24 | Google: Gemini 3.1 Flash Lite (batch) | 1.0M Ultra context (1M+) | $0.13 | Budget | |
| 25 | Google: Gemini 3.1 Flash Lite Preview | 1.0M Ultra context (1M+) | $0.25 | Budget |
Frequently asked questions
How big should the context window be for RAG?
32K is a comfortable minimum for typical knowledge-base RAG. 128K+ lets you include large source documents without summarization. 1M+ enables 'cache the whole codebase' patterns.
Related use cases
AI models for codingAI reasoning modelsAI vision modelslong-context AI models (128K+)AI models for function callingAI models for structured / JSON outputsAI models for agentsfree AI modelsCheapest AI models per tokenmultilingual AI modelsembedding modelssmall / on-device AI modelsAI image generation modelsAI voice / audio modelsAI models for mathAI models for writingopen-source AI modelsAI models for summarizationAI models for translationAI models for enterprise