Benchmark Almanac

Code Review Bench (Martian)

Qodo Extended (product; model unspecified) 65.1% tops its leaderboard
Not in model cards

AI code review tools review pull requests from five large open-source projects. An LLM judge matches their comments to human-verified golden comments.

Offline set of 50 PRs with 173 golden comments, plus a continuously updated online benchmark. It ranks review tools and agents rather than bare models. About 300 GitHub stars.

Made by
Martian
Released
February 2026
Kind
Fixing real repositories
Level
Expert
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
21
Entries on its own leaderboard, 0 models we track

Qodo Extended (product; model unspecified) tops the leaderboard with 65.1%

Best score per entry from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

21 entries. Scroll sideways for the rest.

  1. 65.1 Qodo Extended (product; model unspecified)
  2. 64.9 Cubic v2 (product; model unspecified)
  3. 64.0 Augment (product; model unspecified)
  4. 60.6 Qodo v2 (product; model unspecified)
  5. 58.0 copilot-v2 (product; model unspecified)
  6. 53.6 Gitar (product; model unspecified)
  7. 52.8 Gemini (product; model unspecified)
  8. 52.7 GitLab Duo Code Review (product; model unspecified)
  9. 51.5 Sourcery (product; model unspecified)
  10. 51.2 Greptile v4 (product; model unspecified)
  11. 51.1 CodeRabbit (product; model unspecified)
  12. 50.8 Macroscope (product; model unspecified)
  13. 48.6 Cursor Bugbot (product; model unspecified)
  14. 47.7 Claude Code (CLI) (product; model unspecified)
  15. 45.2 Devin (product; model unspecified)
  16. 41.5 Kodus (product; model unspecified)
  17. 40.9 CodeAnt (product; model unspecified)
  18. 40.5 Baz (product; model unspecified)
  19. 38.9 Claude Code (product; model unspecified)
  20. 18.4 KG (product; model unspecified)
  21. 9.30 Graphite (product; model unspecified)
  1. 1 Qodo Extended (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product qodo-extended-v2 65.1%
  2. 2 Cubic v2 (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product cubic-v2 64.9%
  3. 3 Augment (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product augment 64%
  4. 4 Qodo v2 (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product qodo-v2 60.6%
  5. 5 copilot-v2 (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product copilot-v2 58%
  6. 6 Gitar (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product gitar-v2 53.6%
  7. 7 Gemini (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product gemini-v2 52.8%
  8. 8 GitLab Duo Code Review (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product gitlab 52.7%
  9. 9 Sourcery (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product sourcery 51.5%
  10. 10 Greptile v4 (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product greptile-v4-1 51.2%
  11. 11 CodeRabbit (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product coderabbit 51.1%
  12. 12 Macroscope (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product macroscope 50.8%
  13. 13 Cursor Bugbot (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product bugbot 48.6%
  14. 14 Claude Code (CLI) (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product claude-code 47.7%
  15. 15 Devin (product; model unspecified) core; F2; judge anthropic_claude-sonnet-4-5-20250929; product devin 45.2%
  16. 16 Kodus (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product kodus-v2 41.5%
  17. 17 CodeAnt (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product codeant-v2 40.9%
  18. 18 Baz (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product baz 40.5%
  19. 19 Claude Code (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product claude 38.9%
  20. 20 KG (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product kg 18.4%
  21. 21 Graphite (product; model unspecified) core; F2; judge anthropic_claude-opus-4-5-20251101; product graphite 9.30%

Source: Official Code Review Bench (Martian) leaderboard

#EntryScoreSettingDateSource
1 Qodo Extended (product; model unspecified) 65.1% core; F2; judge anthropic_claude-opus-4-5-20251101; product qodo-extended-v2 Official leaderboard
2 Cubic v2 (product; model unspecified) 64.9% core; F2; judge anthropic_claude-opus-4-5-20251101; product cubic-v2 Official leaderboard
3 Augment (product; model unspecified) 64% core; F2; judge anthropic_claude-opus-4-5-20251101; product augment Official leaderboard
4 Qodo v2 (product; model unspecified) 60.6% core; F2; judge anthropic_claude-opus-4-5-20251101; product qodo-v2 Official leaderboard
5 copilot-v2 (product; model unspecified) 58% core; F2; judge anthropic_claude-opus-4-5-20251101; product copilot-v2 Official leaderboard
6 Gitar (product; model unspecified) 53.6% core; F2; judge anthropic_claude-opus-4-5-20251101; product gitar-v2 Official leaderboard
7 Gemini (product; model unspecified) 52.8% core; F2; judge anthropic_claude-opus-4-5-20251101; product gemini-v2 Official leaderboard
8 GitLab Duo Code Review (product; model unspecified) 52.7% core; F2; judge anthropic_claude-opus-4-5-20251101; product gitlab Official leaderboard
9 Sourcery (product; model unspecified) 51.5% core; F2; judge anthropic_claude-opus-4-5-20251101; product sourcery Official leaderboard
10 Greptile v4 (product; model unspecified) 51.2% core; F2; judge anthropic_claude-opus-4-5-20251101; product greptile-v4-1 Official leaderboard
11 CodeRabbit (product; model unspecified) 51.1% core; F2; judge anthropic_claude-opus-4-5-20251101; product coderabbit Official leaderboard
12 Macroscope (product; model unspecified) 50.8% core; F2; judge anthropic_claude-opus-4-5-20251101; product macroscope Official leaderboard
13 Cursor Bugbot (product; model unspecified) 48.6% core; F2; judge anthropic_claude-opus-4-5-20251101; product bugbot Official leaderboard
14 Claude Code (CLI) (product; model unspecified) 47.7% core; F2; judge anthropic_claude-opus-4-5-20251101; product claude-code Official leaderboard
15 Devin (product; model unspecified) 45.2% core; F2; judge anthropic_claude-sonnet-4-5-20250929; product devin Official leaderboard
16 Kodus (product; model unspecified) 41.5% core; F2; judge anthropic_claude-opus-4-5-20251101; product kodus-v2 Official leaderboard
17 CodeAnt (product; model unspecified) 40.9% core; F2; judge anthropic_claude-opus-4-5-20251101; product codeant-v2 Official leaderboard
18 Baz (product; model unspecified) 40.5% core; F2; judge anthropic_claude-opus-4-5-20251101; product baz Official leaderboard
19 Claude Code (product; model unspecified) 38.9% core; F2; judge anthropic_claude-opus-4-5-20251101; product claude Official leaderboard
20 KG (product; model unspecified) 18.4% core; F2; judge anthropic_claude-opus-4-5-20251101; product kg Official leaderboard
21 Graphite (product; model unspecified) 9.30% core; F2; judge anthropic_claude-opus-4-5-20251101; product graphite Official leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Coding

All 282 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471