# xAI vs DeepSeek

> A closed lab with live X data against an open-weight lab with some of the lowest first-party prices anywhere. Both offer 1M context, with very different trade-offs past it.

Canonical: https://www.subconscious.dev/compare/xai-vs-deepseek · By The Subconscious Team · Updated September 30, 2026

## How they compare

Both are model labs selling their own models through their own APIs, but they sit on opposite sides of the open-weights line. xAI's Grok is closed. Grok 4.6 costs $2 in and $6 out per million under 200K prompt tokens, and the older Grok 4.20 and 4.3 keep a 1M window at $1.25 in and $2.50 out. DeepSeek ships MIT-licensed weights and serves V4.1 Flash at $0.30 in and $1.20 out at peak and V4 Pro at $1.32 in and $3.96 out, both with 1M context, and every off-peak hour costs half. Cache hits run a few cents per million or less.

Long prompts and data location decide a lot here. xAI bills the whole request at double once a prompt hits 200K tokens, which punishes long agent contexts, while DeepSeek's cheap cache hits reward agents that reread long prefixes. DeepSeek's hosted API stores data in China, a hard stop for many enterprises, and it retires and reprices models often. xAI's edge is native Web Search and X Search for current events, plus first-party image, video and audio APIs. Fresh-data agents go to Grok. Cost-first and self-hosted work goes to DeepSeek.

## What each one does

### xAI

xAI sells the Grok models through its own API. Grok 4.6 is the current flagship and xAI tells developers to use it for everything outside audio, image and video, code included. It has a 500K context window and costs $2 in and $6 out per million tokens under 200K prompt tokens. Older Grok 4.20 and 4.3 models keep a 1M window at $1.25 in and $2.50 out, which is aggressive for that capability class.

### DeepSeek

DeepSeek is the Chinese lab whose open-weight models reset price expectations for the whole market. Its API now serves two models, both with 1M context and 384K max output. V4.1 Flash shipped September 10, 2026 with built-in image understanding at $0.30 in and $1.20 out at peak. V4 Pro, generally available since August 13, costs $1.32 in and $3.96 out at peak. Cache hits cost a few cents per million or less, and the weights ship on Hugging Face under an MIT license.

## Which is best, and when

### Choose xAI for

- Agents that need live posts and news from X
- Teams that cannot send data to China-hosted APIs
- Image, video and audio from the same vendor

### Choose DeepSeek for

- Cache-heavy agents on a tight budget
- Batch work scheduled into off-peak hours
- Self-hosting or fine-tuning MIT-licensed weights

## At a glance

| Attribute | xAI | DeepSeek |
|---|---|---|
| Model access | Closed | Open weights (MIT) |
| Flagship models | Grok 4.6, Grok 4.20, grok-build | DeepSeek V4.1 Flash, V4 Pro |
| Speed | ~54 tok/s on Grok 4.6 | ~35 tok/s on V4 Pro |
| Price | $2 in, $6 out (Grok 4.6); 2x past 200K | Off-peak hours at half price |
| Customization | - | Open weights to fine-tune |
| Deployment | First-party API | First-party API, Hugging Face weights |
| Long context | 500K (4.6), 1M (4.20, 4.3) | 1M, 384K max output |

## FAQ

### What is the difference between xAI and DeepSeek?

A closed lab with live X data against an open-weight lab with some of the lowest first-party prices anywhere. Both offer 1M context, with very different trade-offs past it.

### When should I choose xAI over DeepSeek?

Agents that need live posts and news from X; Teams that cannot send data to China-hosted APIs; Image, video and audio from the same vendor.

### When should I choose DeepSeek over xAI?

Cache-heavy agents on a tight budget; Batch work scheduled into off-peak hours; Self-hosting or fine-tuning MIT-licensed weights.

### Is xAI or DeepSeek cheaper?

xAI: $2 in, $6 out (Grok 4.6); 2x past 200K. DeepSeek: Off-peak hours at half price. The cheaper choice depends on the model and workload.

### Which has more context, xAI or DeepSeek?

xAI: 500K (4.6), 1M (4.20, 4.3). DeepSeek: 1M, 384K max output.

## Running long-horizon agents?

If your agents run past 200K tokens, compare both against Subconscious: [Subconscious vs xAI](https://www.subconscious.dev/compare/subconscious-vs-xai.md), [Subconscious vs DeepSeek](https://www.subconscious.dev/compare/subconscious-vs-deepseek.md).

Full profiles: [xAI](https://www.subconscious.dev/providers/xai.md), [DeepSeek](https://www.subconscious.dev/providers/deepseek.md).