BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases8

MirrorCode leaderboard

As of 6 Nov 2026, the top score on MirrorCode is 78.9% by Claude Opus 5.5 (Anthropic), ahead of Claude Fable 5.1 at 73.3%, across 15 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 22 Sep 2026). The model sees a program's behaviour and must write code that mirrors it exactly. MirrorCode is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5.5 holds the top MirrorCode score at 78.9% (Epoch AI Benchmarking Hub, 22 Sep 2026).
  • Claude Fable 5.1 is second on MirrorCode at 73.3% (Epoch AI Benchmarking Hub, 1 Sep 2026).
  • MirrorCode: Reproducing a program's behaviour from its outputs. Run by Epoch AI (Epoch AI Benchmarking Hub).
  • 15 model configurations have a published MirrorCode result on BenchLeader (BenchLeader, 6 Nov 2026).
  • MirrorCode does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Reproducing a program's behaviour from its outputs. Run by Epoch AI.

Published by
Epoch AI Benchmarking Hub
Category
Coding
Index weight
Reference only
Models
15
Data as of
6 Nov 2026

CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.

What the test looks like

The model sees a program's behaviour and must write code that mirrors it exactly.

How it is scored

Percent of programs matched, run by Epoch AI.

What to keep in mind

Very few models measured so far.

Top 15
  1. 1Claude Opus 5.5 (high)78.9%
  2. 2Claude Fable 5.1 (high)73.3%
  3. 3Claude Fable 5 (high)63.9%
  4. 4Gemini 4 Argon (high)58.9%
  5. 5GPT-6 Astra (high)46.7%
  6. 6Gemini 3.7 Flash (high)41.1%
  7. 7Grok 4.6 (xhigh)41.1%
  8. 8Grok 4.5 (high)32.2%
  9. 9Claude Opus 4.7 (high)31.1%
  10. 10GPT-6 Luna (max)26.7%
  11. 11GPT-5.6 Sol (high)20.0%
  12. 12GPT-5.4 (high)15.6%
  13. 13GPT-5.5 (high)10.0%
  14. 14Gemini 3.1 Pro (high)8.9%
  15. 15GPT-5 (high)4.4%
15 of 15
#
1Claude Opus 5.5high78.9%65.82026-09-22
2Claude Fable 5.1high73.3%64.02026-09-01
3Claude Fable 5high63.9%64.32026-06-09
4Gemini 4 Argonhigh58.9%66.02026-09-30
5GPT-6 Astrahigh46.7%67.02026-09-03
6Gemini 3.7 Flashhigh41.1%62.22026-08-13
7Grok 4.6xhigh41.1%57.92026-08-12
8Grok 4.5high32.2%56.92026-07-08
9Claude Opus 4.7high31.1%60.92026-04-16
10GPT-6 Lunamax26.7%56.72026-09-22
11GPT-5.6 Solhigh20.0%63.62026-07-09
12GPT-5.4high15.6%58.82026-03-05
13GPT-5.5high10.0%63.72026-04-23
14Gemini 3.1 Prohigh8.9%60.12026-02-19
15GPT-5high4.4%57.92025-08-07

More coding benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on MirrorCode

Cite as: BenchLeader, “MirrorCode leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/mirrorcode.

MirrorCode: questions

What does MirrorCode measure?
The model sees a program's behaviour and must write code that mirrors it exactly. Scores are reported in percent of tasks solved; higher is better.
What is the top score on MirrorCode?
As of 6 Nov 2026, the highest score on MirrorCode is 78.9% by Claude Opus 5.5 (result dated 22 Sep 2026), ahead of Claude Fable 5.1 at 73.3%. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
Which AI model leads MirrorCode?
Claude Opus 5.5 leads MirrorCode with 78.9% as of 6 Nov 2026 (result dated 22 Sep 2026), ahead of Claude Fable 5.1 at 73.3%.
How many models have MirrorCode results?
15 model configurations have a MirrorCode result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
Who runs MirrorCode and how often is it updated?
MirrorCode is published by Epoch AI Benchmarking Hub. BenchLeader re-reads the published results once a day and records the date each result was published.
Does MirrorCode count toward the BenchLeader Index?
No. MirrorCode is shown for reference but left out of the composite index.