#95 DeepSeek V4 Pro
none
Reached the allocated time limit (300 seconds) without receiving showcase output.
- Cost
- $0.000
- Time
- 300.0s
- Tokens
- 0 tok
DeepSeek
Release: 2026-04-24
Tested on: 2026-05-22 00:38
deepseek/deepseek-v4-pro::none
1.6T total (49B active)MoEOpen source
(high)
(none)
Summary
DeepSeek V4 Pro scores 6.0 on AI BENCHY and ranks #95. It has 8.1 reliability, a 48.3% pass rate, $0.046 total cost, and 13.48s average response time.
What makes DeepSeek V4 Pro unique: It stands out most in Domain specific, where it ranks #3, while Anti-AI Tricks is its weakest area at #17.
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.
6.0
Consistency
8.9
8.1
$0.046
Total Output Tokens
5,347
Total Input Tokens
0
Input Price
$0.435 / 1M
Output Price
$0.870 / 1M
Flaky tests
3
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 03:00 Re-test | 6.8 | 10.0 | $0.256 ↑ | Compare | |
| 2026-07-16 23:18 New test added | 6.9 | 10.0 | $0.096 | Compare | |
| 2026-06-16 15:17 Re-test | 7.2 | 9.9 | $0.034 | Compare | |
| 2026-06-16 14:39 Suite changed | 7.2 | 9.9 | $0.030 | Compare | |
| 2026-06-04 14:24 New test added | 5.7 | 8.5 | $0.025 ↓ | Compare | |
| 2026-05-22 00:38 Suite changed | 6.0 | 8.1 | $0.046 | Current run | |
| 2026-04-29 14:46 Re-test | 6.2 | 7.9 | $0.043 | Compare | |
| 2026-04-24 09:19 Initial run | 3.1 | N/A | $0.009 | Compare |
This run used a different benchmark suite. Keep suite changes in mind when reading historical movement.
Run comparison
| Run | Benchmark coverage | Score | Consistency | Reliability | Tests Correct | Flaky tests | Total Output Tokens | Total Input Tokens | Total Cost | Response Time (avg) |
|---|---|---|---|---|---|---|---|---|---|---|
| 2026-05-22 00:38 · Suite changed | 60/60 attempts | 6.0 | 8.9 | 8.1 | 8/20 | 3 | 5,347 | 0 | $0.046 | 13.48s |
| 2026-07-16 23:18 · New test added | 66/66 attempts | 6.9 | 8.6 | 10.0 | 10/22 | 4 | 35,551 | 148,069 | $0.096 | 11.55s |
| Difference | — | -0.9 | +0.3 | -1.9 | -2 | -1 | -30204 | -148069 | -$0.051 | +1928ms |
Benchmark coverage differs: 60/60 attempts (Target: 3 repeats per test) versus 66/66 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 PrononevsSolar Pro 4mediumDeepSeek V4 PrononevsGLM 5.3lowDeepSeek V4 PrononevsSolar Pro 4lowDeepSeek V4 PrononevsMiMo-V2.5-PromediumDeepSeek V4 PrononevsSeed-2.0-CodemediumDeepSeek V4 PrononevsGemini 3 Flash PreviewnoneDeepSeek V4 PrononevsStep 3.7 FlashhighDeepSeek V4 PrononevsGemma 4 26B A4BmediumFree AvailableDeepSeek V4 PrononevsQwen3.6 FlashmediumDeepSeek V4 PrononevsGLM 5.2none
Category Breakdown
| Category | Score | Consistency | Tests Correct |
|---|---|---|---|
| Anti-AI Tricks | 3.5 | 8.0 | |
| Coding | 5.4 | 6.8 | |
| Combined | 9.5 | 10.0 | |
| Data parsing and extraction | 8.8 | 10.0 | |
| Domain specific | 5.3 | 10.0 | |
| General Intelligence | 4.3 | 9.9 | |
| Instructions following | 6.3 | 10.0 | |
| Puzzle Solving | 7.6 | 7.2 | |
| Tool Calling | 10.0 | 10.0 | |
| Trivia | 3.0 | 10.0 |