AIEvals

Benchmarks

Codeforces (rating)

Competitive-programming strength expressed the way the platform does it: an Elo-style rating instead of a percentage of tasks solved.

Competitive programmingCodeforces (platform); internal harness reported by DeepSeekVendor-reported onlyReported by DeepSeek

How it is run

DeepSeek reports a rating from an internal benchmark built on Codeforces problems; the platform itself does not run models and publishes no comparable table.

What it cannot tell you

A rating is not a percentage and cannot sit in the same column as one. The harness, problem selection and submission settings are not public, so this is a publisher claim with no outside check — DeepSeek reports 3471 for V4.1 Flash against 3348 for V4 Pro.

Why there is no independent table

No independent suite in this dataset runs Codeforces (rating). The numbers below are the ones model publishers report about their own models, which is a different kind of evidence: each publisher chooses its own harness and effort level, so a row is readable on its own and not against the other rows.

Read this with

Benchmarks in the same category, which usually disagree with this one on purpose.

  • IOISame category. Competitive-programming ability on International Olympiad in Informatics problems, with the agent seeing only what a contestant sees: the statement, the sample grader and one sample.

As reported by each publisher

Kept separate per publisher, because the settings are not shared.

DeepSeek

DeepSeek-V4.1-Flash: comparison with closed and open-source models · 10 September 2026

RankModelScore (rating)
2DeepSeek V4 Pro3,348
3DeepSeek V4 Flash3,289
1DeepSeek V4.1 Flash3,471

Ranks compare reported scores in the same published row; ties share a rank.

Codeforces (Rating). A rating, not a percentage: 3471 for V4.1 Flash against 3348 for V4 Pro.

Table 3 of the V4.1 technical report, published with the model release. Every column is the vendor's own run at its maximum reasoning effort, each on that vendor's own harness, so this compares publication claims rather than one shared run. DeepSeek renamed nothing and disclosed no fallback rates; the report instead documents the opposite problem, agents searching for ways to game the harness, and says its visual-agent rows were run with a 512k-token context. The benchmark names are reproduced as DeepSeek wrote them — including rows this site keeps apart, such as HLE with and without tools — and the model columns are renamed to the names used on this site. DeepSeek's table is also republished, unchanged, on the model page at ollama.com; that copy prints 64.0 for the V4.1 row of NL2Repo-Bench where the report prints 65.4, and this site follows the report.

huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/resolve/main/DeepSeek_V41_Tech_Report.pdf

Sources

Every page used for this benchmark, read on 17 October 2026.