RLTasksby ulam.ai
Models / DeepSeek

DeepSeek / MODEL EVIDENCE

DeepSeek V4 Flash

11 benchmarks · 16 observations · 11 sources · released 2026-04-24

Scores at a glance

11 benchmarks · large dot: DeepSeek V4 Flash; small dots: closest models where reported

SimoBench ↗4.429 / 7
3.175.97

DeepSeek-V4-Flash · SimoBench · mean progress / 7 · As published ↗

CritPt ↗7.14%
7.147.14

DeepSeek v4 Flash (max) · CritPt · Accuracy · % ↗

1 more reported result

3.43% · DeepSeek v4 Flash (high) · CritPt · Accuracy · % · Epoch AI · CritPt ↗

LMCA ↗35.88%
2.7868.26

DeepSeek v4 Flash (max) · LMCA · Score · % ↗

DTBench ↗86.4%
41.6598.93

DeepSeek v4 Flash (max) · DTBench · Accuracy · % ↗

SciCode ↗44.91%
44.9144.91

DeepSeek v4 Flash (max) · SciCode · Score · % ↗

1 more reported result

42.01% · DeepSeek v4 Flash (high) · SciCode · Score · % · Epoch AI · SciCode ↗

Closest models by score

Similar positions on shared benchmark leaderboards. Smaller gaps are closer.

15.5 percentile-point gap

5 of 7 shared benchmark comparisons

How similarity is measured

Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.

Filter comparisons to this model →

Sources

Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.

  1. Arena leaderboard — WebDev evaluation ↗

    Arena

  2. SciCode Benchmark | Artificial Analysis ↗

    Artificial Analysis

  3. DTBench — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  4. LMCA — Conceptual Reasoning Index ↗

    Conceptual Reasoning Index

  5. Ale Bench — methodology and published results ↗

    Epoch AI · 2026

  6. CritPt v1 ↗

    Epoch AI

  7. AI benchmarks & capabilities ↗

    Epoch AI

  8. WeirdML ↗

    Håvard Tveit Ihle

  9. DistressBench — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  10. HiL-Bench (Human-in-Loop Benchmark) — evaluation methodology and results ↗

    Scale Labs / SEAL · 2026

  11. SimoBench | Ulam ↗

    Ulam

Names used in source tables

DeepSeek V4 Flash · DeepSeek v4 Flash (high) · DeepSeek v4 Flash (max) · DeepSeek-V4-Flash · deepseek-v4-flash · deepseek-v4-flash-high · deepseek-v4-flash-no-thinking