DeepSeek V4 Pro
Apr 2026 · Open weight · Current · supported evidence
The default cheap-and-good open model for coding. Peak/off-peak API pricing. Not a 60-index flagship.
Intelligence
53.2
AA Index
API price
$0.87 / $1.74
Input / output per 1M
Context
1M
79 tok/s
Capability
Intelligence Index53.2
Coding index68.8
Agentic index49.6
Arena Elo (offset)151
Use it when
- ▸Open-weight coding
- ▸Cost control
- ▸Self-host
Skip if
- –You need the 2026 intelligence cluster (58–63)
Benchmarks
GPQA Diamond Graduate-level science questions designed so Google search is not enough. Still one of the cleanest knowledge/reasoning splits. | — |
|---|---|
Humanity’s Last Exam Expert-written questions across fields. Harder than MMLU; the current differentiator for “does this model actually know things.” | — |
MMLU-Pro Harder, less-saturated successor to MMLU. Classic MMLU is above 90% for every flagship and no longer ranks the field. | — |
SWE-bench Verified 500 human-validated GitHub issues. Score swings 5–15 points by harness — treat vendor numbers as an upper bound. | 80.6% |
SWE-bench Pro Harder, contamination-resistant coding eval. Currently the best public split between “can code” and “can maintain a repo.” | 55.4% |
Terminal-Bench 2.1 End-to-end tasks in a real terminal. Better proxy for coding agents than HumanEval, which is fully saturated. | — |
ARC-AGI-2 Abstract visual puzzles. Rewards generalization over memorization. GPT-5.6 Sol currently leads the published set. | — |
AIME 2025 American Invitational Mathematics Examination. Contest math; reasoning-mode models dominate. | — |
MMMU College-level multimodal understanding across diagrams, charts, and exam figures. | — |
0813 checkpoint is the one on current boards. Official API has peak / off-peak rates.
Modalities
text · tools
~1.6T MoE
Reasoning mode
reasoning
Hybrid and reasoning models spend tokens thinking. That raises GPQA and agents, and also raises latency and bill.