DeepSeek-V4-Pro
DeepSeek moved V4-Pro from preview to general availability across app, web, and API, citing enhanced agent capabilities, reasoning-effort levels, and a peak/off-peak pricing change from Aug 16.
DeepSeek-V4-Pro is available on MiniRouter as . The model page has current pricing, provider routes and API setup.
About the model
- Context: 1M. 384K maximum output per the publisher's spec table
- Off-peak discount: 50% off peak. Peak is 01:00–04:00 and 06:00–10:00 UTC from Aug 16; other hours bill half
- Reasoning effort: low / high / max. Publisher-described effort levels for V4-Pro and V4-Flash
- Weights: Not mentioned. The GA release notes do not mention an open-weights release
Publisher pricing
Pricing below comes from the linked publisher sources and does not represent a MiniRouter route or price.
| Phase | Input / 1M | Output / 1M | Blended / 1M | Effective | Source |
|---|---|---|---|---|---|
| Introductory | $0.435 | $0.87 | $0.54375 | Aug 13, 2026 – Aug 16, 2026 | DeepSeek ↗ |
| Standard · peak | $1.32 | $3.96 | $1.98 | Aug 16, 2026 – open-ended | DeepSeek ↗ |
Publisher-reported benchmarks
Results from the publisher’s announcement, grouped by unit.
| Benchmark | Reported result | Reported by | As of | Source |
|---|---|---|---|---|
| Terminal Bench 2.1 | 87.9 | DeepSeek | Aug 13, 2026 | Announcement ↗ |
| HLE (with tools) | 60.0 | DeepSeek | Aug 13, 2026 | Announcement ↗ |
| DeepSWE | 62.7 | DeepSeek | Aug 13, 2026 | Announcement ↗ |
| Cybergym | 83.3 | DeepSeek | Aug 13, 2026 | Announcement ↗ |
| Toolathlon-Verified | 74.1 | DeepSeek | Aug 13, 2026 | Announcement ↗ |
| GPQA Diamond | 91.2 | DeepSeek | Aug 13, 2026 | Announcement ↗ |
Pricing rises sharply on Aug 16, 2026: peak output moves from $0.87 to $3.96 per 1M tokens, with off-peak hours at half the peak rate.
DeepSeek's own API uses the plain id deepseek-v4-pro; the 0813 suffix is the GA snapshot date.
DeepSeek reports the benchmark figures; MiniRouter has not reproduced them.
Independent measurements
Artificial Analysis evaluations of the first-party API.
DeepSeek V4 Pro 0813 (reasoning max) ranks #28 of 462 model families on their Intelligence Index.
- Intelligence Index
- 36.0
- Composite of their evaluations
- Coding Index
- 68.8
- Coding evaluations only
- Blended price
- $1.98
- USD per 1M tokens, 3:1 input to output, first-party list
- Output speed
- 94
- Median tokens per second
- First answer token
- 22s
- Median seconds, including reasoning
DeepSeek V4 Pro 0813 (reasoning max) on Artificial Analysis ↗
Source: Artificial Analysis ↗, read Sep 21, 2026. Measured on the model's first-party API, not a MiniRouter route.
Keep exploring
Compare published results in the model rankings and coding benchmarks, or read more model releases.
For MiniRouter product updates, visit the changelog.
Sources
- Read at DeepSeek ↗
DeepSeek GA release notes
Published Aug 13, 2026
- Read at Artificial Analysis ↗
Artificial Analysis model page
Published Sep 21, 2026
Editorial reference revised 2026-09-29. Catalog status comes from the generated routing snapshot.