LLM compare · detail

DeepSeek V4 Pro

Provider: DeepSeek

Model type: specialist Budget: open / self-run Use it for: large batches Use it for: coding Use it for: self-hosted stack

batch pick backed by current AA.

Verdict „DeepSeek V4 Pro has IQ 51.5 and input $0.435/M and DeepSWE pass@1 7.5 % in the sources. Consider it for large batches, coding; for enterprise-governance, premium-agents, run a second benchmark before rollout."

Snapshot summary

DeepSeek is a Chinese lab known for aggressive pricing, open-weight direction and models that pressure the big labs on serving cost. DeepSeek V4 Pro is the stronger DeepSeek branch for teams that want a value model with open-weight logic and more control over operations.

In this compare snapshot it is tagged mainly for batch, coding, self-hosted, with price tier free-open and context window není v ověřených datech. The negative side matters too: watch enterprise-governance, premium-agents. Use the numbers below as the hard boundary, because this profile explains the model, it does not replace benchmark provenance.

Figures below, fetched from source artificial_analysis (fetched 2026-06-14T04:00:40Z).

Snapshot figures

self-hosted Budget provider-dependent

51.5 intelligence index artificial_analysis:deepseek-v4-pro

7.5% DeepSWE pass@1 deepswe:deepseek-v4-pro

18.6% DeepSWE pass@4 deepswe:deepseek-v4-pro

What this means

✓ Good for: batch · coding · self-hosted

✗ Not for: enterprise-governance · premium-agents

Source and provenance

Data source: artificial_analysis
Fetched: 2026-06-14T04:00:40Z
Status per metric: price: artificial_analysis:deepseek-v4-pro; IQ: artificial_analysis:deepseek-v4-pro; Elo: curated_seed; Aider: curated_seed; DeepSWE: deepswe:deepseek-v4-pro

This page shows only verified data from the current snapshot.