---
type: "raw_signal"
id: "on-public-cyberbench-deepseek-v4-1-flash-scores-64c908"
slug: "on-public-cyberbench-deepseek-v4-1-flash-scores-64c908"
title: "On public CyberBench , DeepSeek V4.1 Flash scores 73.69% at $0.07/test vs Sol's 77.98% at $1.34/test."
description: "Vals AI’s CyberBench v1.1 leaderboard compares systems on finding and patching security bugs."
retrieval_nugget: "Vals AI’s CyberBench v1.1 leaderboard compares systems on finding and patching security bugs."
observed_at: "2026-10-20"
record_date: "2026-10-20"
date_kind: "observed_at"
why_it_matters: "Vals AI’s CyberBench v1.1 leaderboard compares systems on finding and patching security bugs. DeepSeek V4.1 Flash records 73.69% accuracy at $0.07 per test; GPT-6 Sol records 77.98% at $1.34 per test. The listed results put GPT-6 Sol’s test cost at roughly 19 times DeepSeek V4.1 Flash’s, for a 4.29 percentage-point accuracy difference. These are benchmark results, not evidence of equivalent perfor"
novelty: "notable"
verification_level: "source-linked"
signal_type: "research"
source_platform: "vals.ai"
topics: []
entities: ["Vals AI","CyberBench","DeepSeek V4.1 Flash","GPT-6 Sol"]
related_patterns: []
source_url: "https://www.vals.ai/benchmarks/cyber"
source_urls: ["https://www.vals.ai/benchmarks/cyber"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:on-public-cyberbench-deepseek-v4-1-flash-scores-64c908"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/signals/on-public-cyberbench-deepseek-v4-1-flash-scores-64c908/","markdown":"https://newruntime.com/signals/on-public-cyberbench-deepseek-v4-1-flash-scores-64c908.md","json":"https://newruntime.com/signals/on-public-cyberbench-deepseek-v4-1-flash-scores-64c908.json"}
---

# On public CyberBench , DeepSeek V4.1 Flash scores 73.69% at $0.07/test vs Sol's 77.98% at $1.34/test.

## Retrieval answer

Vals AI’s CyberBench v1.1 leaderboard compares systems on finding and patching security bugs.

Vals AI’s CyberBench v1.1 leaderboard compares systems on finding and patching security bugs. DeepSeek V4.1 Flash records 73.69% accuracy at $0.07 per test; GPT-6 Sol records 77.98% at $1.34 per test. The listed results put GPT-6 Sol’s test cost at roughly 19 times DeepSeek V4.1 Flash’s, for a 4.29 percentage-point accuracy difference. These are benchmark results, not evidence of equivalent performance on production repositories or all security workflows.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)