Skip to main content

Benchmark profile

LiveCodeBench Pass@1 with Chain-of-Thought (LiveCodeBench Pass@1-COT)

This lane contains DeepSeek's LiveCodeBench Pass@1-COT results. The explicit metric and prompting label keeps them separate from generic and version-specific LiveCodeBench rows.

Data verified

How to read this leaderboard

Editorial review by Glevd · 2026-07-15

Compare these rows within the published DeepSeek table. Do not merge them with pass@1 values that use another prompt policy, release window, sampling scheme, or execution configuration.

Operator receipt: 6 sourced rows are currently displayable on this page; the leading published row is DeepSeek V4 Pro (Max) at 93.5%.

Honest limit: All current rows come from one provider report and one model family. They show within-family differences under the reported setup, not a broad cross-provider coding ranking or repository-level software-engineering ability.

Benchmark score on LiveCodeBench Pass@1-COT — July 15, 2026

BenchLM mirrors the published score view for LiveCodeBench Pass@1-COT. DeepSeek V4 Pro (Max) leads the public snapshot at 93.5% , followed by DeepSeek V4 Flash (Max) (91.6%) and DeepSeek V4 Pro (High) (89.8%). BenchLM does not use these results to rank models overall.

6 modelsCodingCurrentDisplay onlyUpdated July 15, 2026

Benchmark score table (6 models)

Score
1
DeepSeek V4 Pro (Max)DeepSeek · Open weight
93.5%
2
DeepSeek V4 Flash (Max)DeepSeek · Open weight
91.6%
3
DeepSeek V4 Pro (High)DeepSeek · Open weight
89.8%
4
DeepSeek V4 Flash (High)DeepSeek · Open weight
88.4%
5
DeepSeek V4 ProDeepSeek · Open weight
56.8%
6
DeepSeek V4 FlashDeepSeek · Open weight
55.2%

The published LiveCodeBench Pass@1-COT snapshot is tightly clustered at the top: DeepSeek V4 Pro (Max) sits at 93.5%, while the third row is only 3.7 points behind. The broader top-10 spread is 38.3 points, so the benchmark still separates strong models even when the leaders cluster.

6 models have been evaluated on LiveCodeBench Pass@1-COT. The benchmark falls in the Coding category. This category carries a 20% weight in BenchLM.ai's overall scoring system. LiveCodeBench Pass@1-COT is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About LiveCodeBench Pass@1-COT

Year

2026

Tasks

DeepSeek-V4 report evaluation window

Format

Pass@1-COT competitive programming results

Difficulty

Competitive programming level

The source reports six reasoning-effort and model variants under a Pass@1-COT column. BenchLM preserves that published metric label and source family without treating the rows as a controlled comparison against generic LiveCodeBench results.

BenchLM freshness & provenance

Version

LiveCodeBench Pass@1-COT 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does LiveCodeBench Pass@1-COT measure?

This lane contains DeepSeek's LiveCodeBench Pass@1-COT results. The explicit metric and prompting label keeps them separate from generic and version-specific LiveCodeBench rows.

Which model scores highest on LiveCodeBench Pass@1-COT?

DeepSeek V4 Pro (Max) by DeepSeek currently leads with a score of 93.5% on LiveCodeBench Pass@1-COT.

How many models are evaluated on LiveCodeBench Pass@1-COT?

6 AI models have been evaluated on LiveCodeBench Pass@1-COT on BenchLM.

Last updated: July 15, 2026 · BenchLM version LiveCodeBench Pass@1-COT 2026

The AI models change fast. We track them for you.

A weekly brief for engineers and researchers covering new models, ranking shifts, and pricing changes.

Free. No spam. Unsubscribe anytime.