DeepSeek V4 Flash 0731 on-premises: what the published scores say, what a token costs, and what sovereignty changes

The independent 2026 benchmark of the three best enterprise AI models: DeepSeek V4 Flash 0731 (284B, 13B active, 1M context, open weights) vs Claude Opus 5 (Anthropic frontier) vs GPT-5.5 (OpenAI frontier). MMLU-Pro, GPQA, AIME, Terminal Bench, cost per token, latency and sovereignty.

By QDNA · Published August 2, 2026 · 10 min read

srcset="/assets/img/blog/en/deepseek-local-dgx-station-spark-vllm.jpg 1200w" sizes="(max-width: 480px) 480px, (max-width: 800px) 800px, 1200px"

Methodology

This benchmark compares the three best enterprise AI models available in August 2026: DeepSeek V4 Flash 0731 (284 billion parameters, 13 billion active per token, 1M context, open weights MIT licence), Claude Opus 5 (Anthropic frontier, 200K-500K context depending on configuration), and GPT-5.5 (OpenAI frontier, 128K-256K context depending on configuration).

The benchmarks used are the official benchmarks published by each vendor, cross-referenced with data from Artificial Analysis Intelligence Index v4.1 (independent methodology with 9 evaluations: GDPval-AA v2, τ³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR).

The cost per token is calculated as total cost of ownership (TCO) over 3 years for local deployment, and as public price for cloud APIs. Latency is measured in real-world conditions (batch=1, 8K context, single-user decode).

Official benchmarks comparison

Scores published by DeepSeek on the DeepSeek-V4-Flash-0731 model card. We do not reproduce scores for Claude Opus 5 and GPT-5.5: their vendors do not publish these same benchmarks, and estimating their values to declare a winner would be worthless. For an up-to-date third-party comparison across models, see Artificial Analysis.

Reading: DeepSeek V4 Flash 0731 is the best on agentic benchmarks (Terminal Bench 2.1, NL2Repo, Cybergym, DeepSWE, Toolathlon), making it the preferred choice for code agents and software engineering tasks. Claude Opus 5 is better on general benchmarks (MMLU-Pro, GPQA, AIME, Humanity's Last Exam, AA-Omniscience), making it the preferred choice for general reasoning and precision. GPT-5.5 is competitive on general benchmarks but less performant on agentic tasks.

Cost per token

The cost per token is calculated as total cost of ownership (TCO) over 3 years for local deployment, and as public price for cloud APIs. The local cost includes hardware acquisition, electricity, integration, maintenance, and training.

Reading: the lowest cost per token is DeepSeek API V4 Flash 0731 at $0.21/M, followed by GPT-5.5 at $2.50/M, Claude Opus 5 at $15/M, and local on DGX Station at €20/M. Local is the most expensive at full load, but it is the cheapest at half-load (sovereignty, no rate-limit, no data transfer, no deactivation risk). The break-even point is at ~1.1 M tokens/day against Claude Opus 5 and at ~3.6 M tokens/day against DeepSeek API.

Latency and performance

The DGX Station column holds our own estimates on that configuration, not certified readings. Third-party API latencies are not published by their vendors: we prefer to write that than to put forward a number.

Reading: serving the model on-premises removes network latency and the vendor rate limit, which is a structural gain rather than a model property. We do not put a figure on the comparison with the APIs: their vendors do not publish these latencies, and they vary with region, load and contract tier. Availability is slightly lower than an API (99.7% against 99.9%), but it does not depend on a third party.

Sovereignty and compliance

Reading: DeepSeek V4 Flash 0731 in local is the only model with open weights, executable on-premises without a revocable licence, no vendor rate-limit, no data transfer outside the Union, and with fine-tuning possible on domain-specific data. Claude Opus 5 and GPT-5.5 are closed APIs that transfer data to Anthropic and OpenAI data centers respectively, with rate-limit and vendor deactivation risk.

Hardware sovereignty: confidentiality and independence through the machine

Choosing local hardware is not only a question of performance or cost per token. It is a sovereignty decision: when the model runs on your own hardware, inside your own infrastructure, you keep full control over data confidentiality, you depend on no third-party vendor, and you are not exposed to rate-limit or deactivation policies.

Confidentiality: prompts, documents injected by RAG, inference logs and generated answers all stay on the company's internal infrastructure. No data crosses the network to a foreign data centre. GDPR compliance is demonstrated by construction rather than by an accumulation of contractual clauses.

Independence: an open-weight model running on company hardware does not fall under a revocable licence. No vendor can suspend access, no rate limit applies, no usage policy can be imposed unilaterally. On July 31, 2026, access to two major AI models was suspended by order of the US authorities for users outside the United States. An open-weight model served on company hardware does not fall under that regime.

Full control: the choice of hardware lets you tune performance, concurrent users, latency, maximum context, without depending on a cloud provider's decisions. You can add memory, change runtime, enable or disable features, and optimise cost per token against your real usage, without a third party's permission.

On hardware sovereignty: local hardware brings full data confidentiality (no transfer outside the Union), complete independence from the vendor (open weights, no revocable licence, no rate limit), and full control over performance (hardware sizing, runtime, configuration). These advantages carry no price tag, but they are decisive for French and European organisations bound by GDPR, HDS or SecNumCloud obligations, or that want to avoid depending on a foreign cloud provider.

Conclusion: which model for which use case

References

• DeepSeek V4 Flash 0731, official Hugging Face repository
• DeepSeek V4 technical report (arXiv 2606.19348)
• Artificial Analysis Intelligence Index v4.1
• Claude Opus 5 pricing (Anthropic)
• OpenAI GPT-5.5 pricing
• Unsloth, DeepSeek V4 documentation
• QDNA sheet: DeepSeek V4
• QDNA sheet: vLLM
• QDNA sheet: DGX Station