BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

SWE Atlas: Codebase QnA leaderboard

As of 6 Nov 2026, the top score on SWE Atlas: Codebase QnA is 63.2% by Claude Opus 5 (Anthropic), ahead of Claude Fable 5.1 at 60.0%, across 22 model configurations with a published result (Scale AI SEAL, result dated 4 Aug 2026). The model must answer questions about how a real codebase works, requiring navigation and reading across files. SWE Atlas: Codebase QnA is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5 holds the top SWE Atlas: Codebase QnA score at 63.2% (Scale AI SEAL, 4 Aug 2026).
  • Claude Fable 5.1 is second on SWE Atlas: Codebase QnA at 60.0% (Scale AI SEAL, 2 Mar 2026).
  • SWE Atlas: Codebase QnA: Answering questions about a large codebase. Scale AI (Scale AI SEAL).
  • 22 model configurations have a published SWE Atlas: Codebase QnA result on BenchLeader (BenchLeader, 6 Nov 2026).
  • SWE Atlas: Codebase QnA does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Answering questions about a large codebase. Scale AI.

Published by
Scale AI SEAL
Category
Coding
Index weight
Reference only
Models
22
Data as of
6 Nov 2026

Scale AI SEAL Leaderboards.

What the test looks like

The model must answer questions about how a real codebase works, requiring navigation and reading across files.

How it is scored

Accuracy, published by Scale AI.

What to keep in mind

Measures code comprehension rather than editing.

Top 15
  1. 1Claude Opus 5 (xhigh)63.2%
  2. 2Claude Fable 5.1 (xhigh)60.0%
  3. 3GPT-6 Astra (xhigh)59.1%
  4. 4Claude Opus 4.8 (xhigh)57.3%
  5. 5GLM 5.248.1%
  6. 6Gemini 3.8 Flash47.0%
  7. 7GPT-5.6 Sol (xhigh)46.0%
  8. 8GPT-5.5 (xhigh)45.4%
  9. 9Muse Spark 1.1 (xhigh)42.2%
  10. 10GPT-5.4 (xhigh)40.8%
  11. 11Claude Opus 4.740.3%
  12. 12Claude Fable 5 (xhigh)39.0%
  13. 13Claude Opus 4.633.3%
  14. 14GPT-5.3 Chat (xhigh)32.6%
  15. 15Claude Sonnet 4.631.2%
22 of 22
#
1Claude Opus 5xhigh63.2%60.22026-08-04
2Claude Fable 5.1xhigh60.0%65.42026-03-02
3GPT-6 Astraxhigh59.1%66.82026-09-09
4Claude Opus 4.8xhigh57.3%–2026-06-08
5GLM 5.248.1%51.52026-06-23
6Gemini 3.8 Flash47.0%54.42026-09-09
7GPT-5.6 Solxhigh46.0%64.22026-03-02
8GPT-5.5xhigh45.4%64.32026-05-07
9Muse Spark 1.1xhigh42.2%59.52026-07-09
10GPT-5.4xhigh40.8%63.22026-03-09
11Claude Opus 4.740.3%63.12026-06-18
12Claude Fable 5xhigh39.0%–2026-03-02
13Claude Opus 4.633.3%62.32026-02-25
14GPT-5.3 Chatxhigh32.6%–2026-02-25
15Claude Sonnet 4.631.2%58.22026-02-25
16DeepSeek V4 Pro27.1%53.22026-06-18
17Muse Spark24.2%64.12026-04-08
18GLM-520.5%54.82026-02-25
19Gemini 3.1 Pro13.5%60.52026-02-25
20Kimi K2.513.1%54.22026-02-25
21MiniMax M2.510.3%53.62026-02-25
22Gemini 3 Flash8.2%56.92026-02-25

More coding benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on SWE Atlas: Codebase QnA

Cite as: BenchLeader, “SWE Atlas: Codebase QnA leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/scale_sweatlas_qna.

SWE Atlas: Codebase QnA: questions

What does SWE Atlas: Codebase QnA measure?
The model must answer questions about how a real codebase works, requiring navigation and reading across files. Scores are reported in percent of tasks solved; higher is better.
What is the top score on SWE Atlas: Codebase QnA?
As of 6 Nov 2026, the highest score on SWE Atlas: Codebase QnA is 63.2% by Claude Opus 5 (result dated 4 Aug 2026), ahead of Claude Fable 5.1 at 60.0%. Scores come from Scale AI SEAL and are re-read daily.
Which AI model leads SWE Atlas: Codebase QnA?
Claude Opus 5 leads SWE Atlas: Codebase QnA with 63.2% as of 6 Nov 2026 (result dated 4 Aug 2026), ahead of Claude Fable 5.1 at 60.0%.
How many models have SWE Atlas: Codebase QnA results?
22 model configurations have a SWE Atlas: Codebase QnA result on BenchLeader, all taken from Scale AI SEAL.
Who runs SWE Atlas: Codebase QnA and how often is it updated?
SWE Atlas: Codebase QnA is published by Scale AI SEAL. BenchLeader re-reads the published results once a day and records the date each result was published.
Does SWE Atlas: Codebase QnA count toward the BenchLeader Index?
No. SWE Atlas: Codebase QnA is shown for reference but left out of the composite index.