Model comparison
Qwen3 (2507 update) vs DeepSeek-V3-0324
- Input
- $0.09
- Output
- $0.10
- Context
- 262K (extendable to 1M)
- Released
- 2025-07
- Input
- $0.27 / 1M tokens (cache miss)
- Output
- $1.10 / 1M tokens
- Context
- 128K
- Released
- 2025-03-24
Qwen3 (2507 update) scores higher on 2 of the 2 benchmarks both models report; Qwen3 (2507 update) is cheaper on input tokens ($0.1 vs $0.3 per 1M, 3×); Qwen3 (2507 update) has the larger context window (262K vs 128K).
Specs and pricing side by side
| Spec | Qwen3 (2507 update) | DeepSeek-V3-0324 |
|---|---|---|
| Maker | Alibaba (Qwen) | DeepSeek |
| Released | 2025-07 | 2025-03-24 |
| Context window | 262K (extendable to 1M) | 128K |
| Max output | 81,920 tokens (Thinking) | 8K |
| Input price | $0.09 | $0.27 / 1M tokens (cache miss) |
| Output price | $0.10 | $1.10 / 1M tokens |
| Parameters | 235B total · 22B active (also 30B-A3B) | 671B total / 37B active (685B checkpoint) |
| License | Apache-2.0 | MIT |
| Weights | Open weights | Open weightssame |
| Modalities | Text | Textsame |
| Knowledge cutoff | not published | July 2024 |
Benchmarks both models report
| Benchmark | Qwen3 (2507 update) | DeepSeek-V3-0324 |
|---|---|---|
| MMLU-Pro | 84.4% | 81.2% |
| GPQA | 81.1% | 68.4% |
Scores are as published by each maker or its cited source; harness, reasoning effort and date can differ between rows, so treat small gaps as ties. Sources are listed on each model page.
Which one should you pick?
Pick Qwen3 (2507 update) if…
- You care about MMLU-Pro, GPQA (it scores higher on 2 shared benchmarks).
- Cost matters: input tokens cost 3× less ($0.1 vs $0.3 per 1M).
- You need the longer context (262K vs 128K).
- Your use case is: self-hosted assistants and agents where an open Apache 2.0 license and on-prem control matter.
Pick DeepSeek-V3-0324 if…
- Your use case is: self-hosted general-purpose chat and assistant deployments where open weights and MIT licensing matter.
FAQ
Which is better, Qwen3 (2507 update) or DeepSeek-V3-0324?
Qwen3 (2507 update) scores higher on 2 of the 2 benchmarks both models report; Qwen3 (2507 update) is cheaper on input tokens ($0.1 vs $0.3 per 1M, 3×); Qwen3 (2507 update) has the larger context window (262K vs 128K). The better pick depends on whether you value benchmark scores, price per token, context length or openness for your workload.
How much cheaper is Qwen3 (2507 update)?
Qwen3 (2507 update): $0.09 input, $0.10 output. DeepSeek-V3-0324: $0.27 / 1M tokens (cache miss) input, $1.10 / 1M tokens output (list API prices as published by each maker).
What are the context windows of Qwen3 (2507 update) and DeepSeek-V3-0324?
Qwen3 (2507 update) offers 262K (extendable to 1M) with up to 81,920 tokens (Thinking) output; DeepSeek-V3-0324 offers 128K with up to 8K output.
Is Qwen3 (2507 update) newer than DeepSeek-V3-0324?
Qwen3 (2507 update) was released 2025-07; DeepSeek-V3-0324 was released 2025-03-24.