Reasoning · Benchmark profile
A million-token CorpusQA long-context question-answering benchmark reported in DeepSeek-V4 model evaluations.
Data verified 21 Jul 2026 · Methodology 1.6.0
Benchmark score on CorpusQA 1M
DeepSeek V4 Pro (Max) leads at 62%, followed by DeepSeek V4 Flash (Max) (60.5%) and DeepSeek V4 Flash (High) (59.3%).
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
6 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| DeepSeek V4 Pro (Max) | #1 | DeepSeek | 62% |
| DeepSeek V4 Flash (Max) | #2 | DeepSeek | 60.5% |
| DeepSeek V4 Flash (High) | #3 | DeepSeek | 59.3% |
| DeepSeek V4 Pro (High) | #4 | DeepSeek | 56.5% |
| DeepSeek V4 Pro | #5 | DeepSeek | 35.6% |
| DeepSeek V4 Flash | #6 | DeepSeek | 15.5% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | DeepSeek V4 Pro (Max) deepseek-v4-pro-max | DeepSeek | open | Estimated reference | 62% |
| #2 | DeepSeek V4 Flash (Max) deepseek-v4-flash-max | DeepSeek | open | Estimated reference | 60.5% |
| #3 | DeepSeek V4 Flash (High) deepseek-v4-flash-high | DeepSeek | open | Estimated reference | 59.3% |
| #4 | DeepSeek V4 Pro (High) deepseek-v4-pro-high | DeepSeek | open | Estimated reference | 56.5% |
| #5 | DeepSeek V4 Pro deepseek-v4-pro | DeepSeek | open | Reference only | 35.6% |
| #6 | DeepSeek V4 Flash deepseek-v4-flash | DeepSeek | open | Estimated reference | 15.5% |
The top of this snapshot is led by DeepSeek V4 Pro (Max) at 62%; third place is 2.7 points behind. The top-6 spread is 46.5 points.
About CorpusQA 1M
A million-token CorpusQA long-context question-answering benchmark reported in DeepSeek-V4 model evaluations. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A million-token CorpusQA long-context question-answering benchmark reported in DeepSeek-V4 model evaluations.
DeepSeek V4 Pro (Max) by DeepSeek currently leads with 62%.
6 models in the LuminaBench cohort have a qualifying score on this benchmark.
Related
