DeepSeek V3.2 vs Grok 4
Benchmarks, API pricing and specs, head to head. Data updated 2026-06-10.
DeepSeek · Sep 2025
The sparse-attention release that halved DeepSeek's already-lowest-in-class prices while keeping GPT-class quality.
xAI · Jul 2025
xAI's mid-2025 flagship, trained on the 200K-GPU Colossus cluster — a math and science reasoning standout of its generation.
The verdict
Grok 4 wins 3 of the 3 benchmarks these models share, against 0 for DeepSeek V3.2. On price the gap is dramatic: DeepSeek V3.2 works out roughly 19x cheaper per blended million tokens. Grok 4 also takes 256K of context versus 128K for DeepSeek V3.2. And DeepSeek V3.2 is open-weights (MIT), so it can be self-hosted — a structural advantage if data control or vendor independence matters.
Benchmark head-to-head 0–3
Specs & pricing
| DeepSeek V3.2 | Grok 4 | |
|---|---|---|
| modhub Index | 68.4 | — |
| Input price / 1M | $0.28 | $3 |
| Output price / 1M | $0.42 | $15 |
| Context window | 128K | 256K |
| Max output | 64K | — |
| Open weights | yes (MIT) | no |
| Reasoning model | yes | yes |
| Multimodal input | text | text, image |
| Knowledge cutoff | Jul 2025 | Jul 2025 |
| Released | Sep 2025 | Jul 2025 |
| Example monthly cost* | $3.43 | $52.50 |
* 10M input + 1.5M output tokens per month at list prices, no caching. Green = better value on that row.
Frequently asked questions
- Which is better, DeepSeek V3.2 or Grok 4?
- Grok 4 wins 3 of the 3 benchmarks these models share, against 0 for DeepSeek V3.2. On price the gap is dramatic: DeepSeek V3.2 works out roughly 19x cheaper per blended million tokens. Grok 4 also takes 256K of context versus 128K for DeepSeek V3.2. And DeepSeek V3.2 is open-weights (MIT), so it can be self-hosted — a structural advantage if data control or vendor independence matters.
- Which is cheaper, DeepSeek V3.2 or Grok 4?
- DeepSeek V3.2 costs $0.28/$0.42 per million input/output tokens, while Grok 4 costs $3/$15. For a typical workload of 10M input and 1.5M output tokens per month, that's $3.43 versus $52.50.
- Which model is better for coding, DeepSeek V3.2 or Grok 4?
- We don't yet track SWE-bench Verified results for both models; check their individual pages for coding-related scores.