BenchLeader

SWE-Bench Pro V2 leaderboard

As of 16 Nov 2026, the top score on SWE-Bench Pro V2 is 99.4% by Claude Opus 5 (Anthropic), ahead of Gemini 4 Argon at 99.2%, across 18 model configurations with a published result (Scale AI SEAL, result dated 7 Oct 2026). SWE-Bench Pro's refreshed public split (September 2026): 642 tasks in 11 repositories, web tools off, every patch re-graded on a clean image. Scale AI. SWE-Bench Pro V2 is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 16 Nov 2026)

  • Claude Opus 5 holds the top SWE-Bench Pro V2 score at 99.4% (Scale AI SEAL, 7 Oct 2026).
  • Gemini 4 Argon is second on SWE-Bench Pro V2 at 99.2% (Scale AI SEAL, 14 Oct 2026).
  • SWE-Bench Pro V2: SWE-Bench Pro's refreshed public split (September 2026): 642 tasks in 11 repositories, web tools off, every patch re-graded on a clean image. Scale AI (Scale AI SEAL).
  • 18 model configurations have a published SWE-Bench Pro V2 result on BenchLeader (BenchLeader, 16 Nov 2026).
  • SWE-Bench Pro V2 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

SWE-Bench Pro's refreshed public split (September 2026): 642 tasks in 11 repositories, web tools off, every patch re-graded on a clean image. Scale AI.

Published by
Scale AI SEAL
Category
Coding
Index weight
Reference only
Models
18
Data as of
16 Nov 2026
Newest result
6 Nov 2026

Scale AI SEAL Leaderboards.

18 of 18
#
1Claude Opus 5xhigh99.4%60.12026-10-07
2Gemini 4 Argonhigh99.2%65.32026-10-14
3Claude Fable 5.1high99.1%63.72026-10-07
4Kimi K3max97.7%60.22026-10-07
5Inkling 2 Small97.7%57.92026-11-06
6GPT-6 Astrahigh96.9%66.32026-10-07
7Muse Wonderxhigh96.7%–2026-10-14
8MiniMax M3.1 Flashxhigh96.6%–2026-11-02
9GPT-6.1 Solxhigh96.1%63.62026-10-07
10Claude Sonnet 5.5xhigh95.8%62.52026-10-12
11GLM 5.3max95.6%58.82026-10-07
12GPT-5.6 Solxhigh95.5%63.92026-10-07
13Muse Spark Openxhigh95.0%53.52026-10-16
14Gemini 3.8 Flashhigh94.9%59.42026-10-07
15Claude Sonnet 5xhigh93.2%55.42026-10-07
16GPT-5.6 Terraxhigh92.4%60.12026-10-07
17Inklingxhigh89.9%50.02026-10-07
18Claude Haiku 5.5high87.2%55.32026-10-08

AI coding benchmark leaderboard · All LLM benchmarks

Cite as: BenchLeader, “SWE-Bench Pro V2 leaderboard”, data as of 16 Nov 2026, https://www.benchleader.com/benchmarks/scale_swe_bench_pro_v2.

SWE-Bench Pro V2: questions

What does SWE-Bench Pro V2 measure?
SWE-Bench Pro's refreshed public split (September 2026): 642 tasks in 11 repositories, web tools off, every patch re-graded on a clean image. Scale AI. Scores are reported in percent of tasks solved; higher is better.
What is the top score on SWE-Bench Pro V2?
As of 16 Nov 2026, the highest score on SWE-Bench Pro V2 is 99.4% by Claude Opus 5 (result dated 7 Oct 2026), ahead of Gemini 4 Argon at 99.2%. Scores come from Scale AI SEAL and are re-read daily.
Which AI model leads SWE-Bench Pro V2?
Claude Opus 5 leads SWE-Bench Pro V2 with 99.4% as of 16 Nov 2026 (result dated 7 Oct 2026), ahead of Gemini 4 Argon at 99.2%.
How many models have SWE-Bench Pro V2 results?
18 model configurations have a SWE-Bench Pro V2 result on BenchLeader, all taken from Scale AI SEAL.
Who runs SWE-Bench Pro V2 and how often is it updated?
SWE-Bench Pro V2 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-Bench Pro V2 count toward the BenchLeader Index?
No. SWE-Bench Pro V2 is shown for reference but left out of the composite index.