DeepSeek
DeepSeek V4 Flash 0731
Launched 2026-07-31 · Knowledge up to 2025-05 · Listed on OpenRouter 2026-07-31
Prices updated . Benchmarks updated .
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....
You send / 1M
$0.060
tokens in
It replies / 1M
$0.180
tokens out
Context
1M
tokens it can read at once
Typical request
$0.0002
1,000 words in, 400 out
Benchmarks
| Benchmark | Score | Kind | Date | Source |
|---|---|---|---|---|
| Epoch Capabilities Index | 154.9 | Independent | 2026-07-31 | Epoch AI |
| GPQA DiamondMax effort | 91.0% | Independent | 2026-08-02 | Epoch AI |
| OTIS Mock AIMEMax effort | 94.4% | Independent | 2026-08-02 | Epoch AI |
| SimpleQA VerifiedMax effort | 33.6% | Independent | 2026-08-27 | Epoch AI |
| ARC-AGI-2Max effort | 61.4% | Independent | 2026-07-31 | ARC Prize, republished by Epoch AI |
| CritPtMax effort | 16.6% | Independent | 2026-07-31 | CritPt, republished by Epoch AI |
Rankings on this site use the Epoch Capabilities Index, an independent composite. Lab-reported rows are the lab's own figures and are kept separate.
Data privacy
No training or retention: Yes. At least one provider serves it with zero data retention.
| Provider | You send / 1M | It replies / 1M | Uptime | Privacy policy |
|---|---|---|---|---|
| Relace · fp4 | $0.070 | $0.140 | 100% | Read |
| DeepInfra · fp8 | $0.048 | $0.144 | 97% | Read |
| DeepInfra · fp8 | $0.060 | $0.180 | 100% | Read |
| Wafer · nvfp4 | $0.140 | $0.230 | 100% | Read |
| DigitalOcean | $0.119 | $0.238 | 100% | Read |
| BaseTen · fp8 | $0.130 | $0.260 | 100% | Read |
| BaseTen · fp32 | $0.130 | $0.260 | 100% | Read |
| Venice | $0.130 | $0.260 | 100% | Read |
| DeepInfra · fp8 | $0.090 | $0.270 | 100% | Read |
| CoreWeave · fp8 | $0.130 | $0.270 | 100% | Read |
| Cohere · fp8 | $0.140 | $0.280 | 100% | Read |
| Parasail · fp8 | $0.140 | $0.280 | 99% | Read |
| Wafer · nvfp4 | $0.043 | $0.300 | 100% | Read |
| Sail Research · fp4 | $0.150 | $0.300 | 99% | Read |
| Sail Research · fp4 | $0.150 | $0.300 | 99% | Read |
| Reka | $0.100 | $0.420 | 100% | Read |
| Inceptron · fp4 | $0.130 | $0.495 | 100% | Read |
| Mancer 2 · fp8 | $0.150 | $0.500 | 98% | Read |
| SiliconFlow · fp8 | $0.220 | $0.660 | 99% | Read |
| Novita · fp8 | $0.409 | $1.23 | 100% | Read |
| PrimeIntellect · fp4 | $0.440 | $1.32 | 100% | Read |
| Cloudflare | $0.440 | $1.32 | 100% | Read |
Based on OpenRouter and Vercel AI Gateway data. Confirm terms with the provider before sending sensitive data.
What it can do
- Reads: text
- Writes: text
- Weights: Open weights
- Reasoning: Yes
- Tool calling: Yes
- Longest reply: 393K tokens
- Repeated text is billed at $0.015 per 1M (prompt caching)
Compare with
Appears in
Estimate your monthly cost with DeepSeek V4 Flash 0731
Sources and licenses
- OpenRouter: Prices, context windows, modalities and the model list come from the OpenRouter public API. Data: OpenRouter.
- Vercel AI Gateway: The no-training flag comes from the Vercel AI Gateway model list.
- models.dev: Lab launch dates, open-weights flags and knowledge cutoffs come from models.dev. Copyright (c) 2025 models.dev, MIT License.
- Epoch AI: Independent benchmark scores and the Epoch Capabilities Index (ECI) come from Epoch AI's Benchmarking Hub, used under CC BY 4.0. Epoch AI, Capabilities & benchmarking, epoch.ai/benchmarks. Scores Epoch marks as external keep their original license, and the source is named on each score.
- LMArena: Arena Elo (text, webdev and vision) comes from the LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), CC BY 4.0, leaderboard date 2026-10-28.
- Lab-reported scores, when an editor has added them, are the lab's own numbers. Each one links to the source and is labeled lab-reported. They are not used to rank the lists.
- What a score measures: the benchmark guide.
