BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

DeepSWE leaderboard

As of 6 Nov 2026, the top score on DeepSWE is 74.1% by GPT-6 Astra (OpenAI), ahead of Gemini 3.8 Flash at 73.8%, across 69 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 3 Sep 2026). SWE-style repository tasks run in an agent loop, with the reasoning effort noted per run. DeepSWE is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • GPT-6 Astra holds the top DeepSWE v1.1 score at 74.1% (Epoch AI Benchmarking Hub, 3 Sep 2026).
  • Gemini 3.8 Flash is second on DeepSWE v1.1 at 73.8% (Epoch AI Benchmarking Hub, 2 Sep 2026).
  • DeepSWE v1.1: Software-engineering tasks in an agent harness with reasoning effort recorded (Epoch AI Benchmarking Hub).
  • 69 model configurations have a published DeepSWE v1.1 result on BenchLeader (BenchLeader, 6 Nov 2026).
  • DeepSWE v1.1 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Software-engineering tasks in an agent harness with reasoning effort recorded.

Published by
Epoch AI Benchmarking Hub
Category
Coding
Index weight
Reference only
Models
69
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

SWE-style repository tasks run in an agent loop, with the reasoning effort noted per run.

How it is scored

Pass@1, as published.

What to keep in mind

Harness-dependent.

Top 15
  1. 1GPT-6 Astra (xhigh)74.1%
  2. 2Gemini 3.8 Flash (high)73.8%
  3. 3Claude Opus 5 (max)73.7%
  4. 4GPT-6 Astra (max)73.2%
  5. 5GPT-6 Astra (high)73.2%
  6. 6Claude Opus 5 (xhigh)73.2%
  7. 7Claude Opus 5 (high)72.8%
  8. 8GPT-6 Astra (medium)72.8%
  9. 9GPT-5.6 Sol (max)72.7%
  10. 10Gemini 3.8 Flash (medium)71.0%
  11. 11GPT-5.6 Sol (xhigh)70.7%
  12. 12Claude Fable 5 (xhigh)69.9%
  13. 13Claude Fable 5 (max)69.7%
  14. 14GPT-5.6 Terra (max)69.6%
  15. 15GPT-5.6 Sol (high)69.4%
69 of 69
#
1GPT-6 Astraxhigh74.1%66.82026-09-03
2Gemini 3.8 Flashhigh73.8%62.02026-09-02
3Claude Opus 5max73.7%65.42026-07-24
4GPT-6 Astramax73.2%67.82026-09-03
5GPT-6 Astrahigh73.2%67.02026-09-03
6Claude Opus 5xhigh73.2%60.22026-07-24
7Claude Opus 5high72.8%64.32026-07-24
8GPT-6 Astramedium72.8%65.12026-09-03
9GPT-5.6 Solmax72.7%65.42026-07-09
10Gemini 3.8 Flashmedium71.0%60.72026-09-02
11GPT-5.6 Solxhigh70.7%64.22026-07-09
12Claude Fable 5xhigh69.9%–2026-06-09
13Claude Fable 5max69.7%65.32026-06-09
14GPT-5.6 Terramax69.6%60.72026-07-09
15GPT-5.6 Solhigh69.4%63.62026-07-09
16GLM 5.3max69.0%59.02026-08-14
17Claude Opus 5medium68.9%57.42026-07-24
18Claude Fable 5high68.6%64.32026-06-09
19Kimi K3max68.5%62.12026-07-16
20Grok 4.6medium67.5%59.62026-08-12
21GPT-5.6 Lunamax67.2%56.32026-07-09
22GPT-5.5xhigh67.0%64.32026-04-23
23GPT-6 Astralow67.0%62.72026-09-03
24Grok 4.6xhigh66.7%57.92026-08-12
25Gemini 3.7 Flashmedium65.5%60.32026-08-13
26Claude Fable 5medium65.4%–2026-06-09
27Gemini 3.7 Flashhigh65.3%62.22026-08-13
28Grok 4.6high65.2%60.52026-08-12
29GPT-5.5high64.4%63.72026-04-23
30GLM 5.3 Flashmax63.4%54.22026-08-20
31GPT-5.6 Solmedium61.1%61.52026-07-09
32GPT-5.6 Terraxhigh60.2%60.32026-07-09
33Claude Fable 5low59.6%–2026-06-09
34Claude Opus 4.8max59.0%60.22026-05-28
35Claude Opus 5low58.1%57.02026-07-24
36Qwen3.8 Maxxhigh57.5%57.12026-08-02
37GPT-5.6 Lunaxhigh56.9%58.52026-07-09
38Muse Spark 1.2xhigh54.9%61.72026-08-05
39Claude Opus 4.8xhigh54.4%–2026-05-28
40GPT-5.5medium54.0%61.02026-04-23
41Claude Sonnet 5max53.9%53.82026-06-30
42Gemini 3.7 Flashlow53.8%57.52026-08-13
43GPT-5.6 Terrahigh53.8%57.02026-07-09
44Grok 4.5high53.8%56.92026-07-08
45Muse Spark 1.153.3%62.22026-07-09
46GPT-5.4xhigh51.8%63.22026-03-05
47Claude Opus 4.8high51.8%60.62026-05-28
48Claude Sonnet 5xhigh49.7%55.62026-06-30
49Claude Opus 4.8medium48.7%–2026-05-28
50Claude Sonnet 5high48.2%59.32026-06-30
51Gemini 3.6 Flashhigh46.7%61.22026-07-21
52GPT-5.6 Sollow45.4%60.42026-07-09
53GPT-5.6 Lunahigh44.3%54.82026-07-09
54GLM 5.2max43.8%59.52026-06-16
55Grok 4.6low41.6%55.22026-08-12
56Claude Opus 4.8low40.8%–2026-05-28
57Claude Sonnet 5medium39.8%–2026-06-30
58Gemini 3.5 Flashmedium37.4%61.92026-05-19
59GLM 5.2high36.3%–2026-06-16
60Gemini 3.5 Flashhigh36.1%62.52026-05-19

More coding benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on DeepSWE v1.1

Cite as: BenchLeader, “DeepSWE v1.1 leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/deepswe.

DeepSWE v1.1: questions

What does DeepSWE v1.1 measure?
SWE-style repository tasks run in an agent loop, with the reasoning effort noted per run. Scores are reported in percent of tasks solved; higher is better.
What is the top score on DeepSWE?
As of 6 Nov 2026, the highest score on DeepSWE v1.1 is 74.1% by GPT-6 Astra (result dated 3 Sep 2026), ahead of Gemini 3.8 Flash at 73.8%. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
Which AI model leads DeepSWE v1.1?
GPT-6 Astra leads DeepSWE v1.1 with 74.1% as of 6 Nov 2026 (result dated 3 Sep 2026), ahead of Gemini 3.8 Flash at 73.8%.
How many models have DeepSWE v1.1 results?
69 model configurations have a DeepSWE v1.1 result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
Who runs DeepSWE v1.1 and how often is it updated?
DeepSWE v1.1 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 DeepSWE v1.1 count toward the BenchLeader Index?
No. DeepSWE v1.1 is shown for reference but left out of the composite index.