#51 DeepSeek V4 Pro
high- Cost
- $0.023
- Time
- 257.6s
- Tokens
- 14,870 tok
DeepSeek
Release: 2026-04-24
Tested on: 2026-04-29 14:47
deepseek/deepseek-v4-pro::high
1.6T total (49B active)MoEOpen source
(high)
(none)
Summary
DeepSeek V4 Pro 0423 scores 7.5 on AI BENCHY and ranks #51. It has 9.3 reliability, a 73.2% pass rate, $0.209 total cost, and 60.03s average response time.
What makes DeepSeek V4 Pro 0423 unique: Its total benchmark cost is unusually low for its score range.
Model facts
Researched on 2026-08-12
- Parameters
- 1.6T total (49B active)
- Architecture
- MoE
- Availability
- Open source
- License
- MIT
The V4 family card documents the Pro configuration.
7.5
Consistency
8.0
9.3
$0.209
Total Output Tokens
62,125
Total Input Tokens
0
Input Price
$0.435 / 1M
Output Price
$0.870 / 1M
Cache Read Price
N/A
Cache Write Price
N/A
Cache prices apply to input tokens. Reads reuse cached prompts; writes store them and can cost extra. Output tokens use the output price.
Flaky tests
4
Flaky tests had mixed outcomes across runs (at least one pass and one fail).
Hamster playing table tennis
Prompt: Create a detailed SVG illustration of a hamster playing table tennis.
Run history
| Tested on | Score | Reliability | Tests Correct | Total Cost | Compare |
|---|---|---|---|---|---|
| 2026-08-14 03:17 Re-test | 7.7 | 10.0 | $0.555 ↑ | Compare | |
| 2026-07-16 23:19 New test added | 7.7 | 10.0 | $0.200 | Compare | |
| 2026-06-16 15:31 Re-test | 7.6 | 9.3 | $0.157 | Compare | |
| 2026-06-16 14:47 Suite changed | 8.1 | 9.6 | $0.098 | Compare | |
| 2026-06-04 14:38 New test added | 6.0 | 8.9 | $0.079 ↓ | Compare | |
| 2026-05-22 00:54 Suite changed | 6.6 | 9.0 | $0.212 | Compare | |
| 2026-04-29 14:47 Re-test | 7.5 | 9.3 | $0.209 | Current run | |
| 2026-04-26 10:50 Re-test | 7.5 | 8.4 | $0.201 | Compare | |
| 2026-04-25 21:53 Initial run | 8.2 | N/A | $0.329 | Compare |
Run comparison
| Run | Benchmark coverage | Score | Consistency | Reliability | Tests Correct | Flaky tests | Total Output Tokens | Total Input Tokens | Total Cost | Response Time (avg) |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-04-29 14:47 · Re-test | 54/54 attempts | 7.5 | 8.0 | 9.3 | 11/18 | 4 | 62,125 | 0 | $0.209 | 60.03s |
| 2026-06-16 15:31 · Re-test | 63/63 attempts | 7.6 | 7.0 | 9.3 | 9/21 | 8 | 165,485 | 38,726 | $0.157 | 77.20s |
| Difference | — | -0.1 | +1.0 | 0.0 | +2 | -4 | -103360 | -38726 | +$0.052 | -17165ms |
Benchmark coverage differs: 54/54 attempts (Target: 3 repeats per test) versus 63/63 attempts (Target: 3 repeats per test). Totals and repeat-sensitive metrics are not directly comparable.
These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.
Charts
Choose the first model, then click a second model to open a side-by-side page.
Top Models by Score
Score vs Total Cost
Response Time (avg)
Score vs Response Time (avg)
Total Output Tokens
Score vs Total Output Tokens
Quick Compare
DeepSeek V4 Pro 0423highvsSeed-2.0-CodelowDeepSeek V4 Pro 0423highvsClaude Opus 4.7mediumDeepSeek V4 Pro 0423highvsSeed-2.0-LitemediumDeepSeek V4 Pro 0423highvsDeepSeek V4 Flash 0423highDeepSeek V4 Pro 0423highvsGPT-5.6 TerramediumDeepSeek V4 Pro 0423highvsDeepSeek V4.1 FlashlowDeepSeek V4 Pro 0423highvsGLM 5.2mediumDeepSeek V4 Pro 0423highvsGPT-5.6 LunahighDeepSeek V4 Pro 0423highvsGrok 4.5highDeepSeek V4 Pro 0423highvsNemotron 3 UltramediumFree Available
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Anti-AI Tricks | 8.3 | 10.0 | |
| Coding | 3.0 | 0.0 | |
| Combined | 10.0 | 10.0 | |
| Data parsing and extraction | 10.0 | 10.0 | |
| Domain specific | 3.2 | 6.9 | |
| General Intelligence | 6.1 | 3.1 | |
| Instructions following | 10.0 | 10.0 | |
| Puzzle Solving | 8.0 | 7.0 | |
| Tool Calling | 10.0 | 10.0 |