BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases4

GameBench 2 Overall leaderboard

As of 2 Nov 2026, the top score on GameBench 2 Overall is 89.7 by GPT-6 Astra (OpenAI), ahead of Claude Opus 5.5 at 88.8, across 48 model configurations with a published result (DuelLab GameBench 2, result dated 26 Oct 2026). DuelLab's overall GameBench 2 score: models write game-playing programs, the programs play each other, and the rating comes from match results. Shown only for models DuelLab ranks overall. GameBench 2 Overall is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 2 Nov 2026)

  • GPT-6 Astra holds the top GameBench 2 Overall score at 89.7 (DuelLab GameBench 2, 26 Oct 2026).
  • Claude Opus 5.5 is second on GameBench 2 Overall at 88.8 (DuelLab GameBench 2, 26 Oct 2026).
  • GameBench 2 Overall: DuelLab's overall GameBench 2 score: models write game-playing programs, the programs play each other, and the rating comes from match results. Shown only for models DuelLab ranks overall (DuelLab GameBench 2).
  • 48 model configurations have a published GameBench 2 Overall result on BenchLeader (BenchLeader, 2 Nov 2026).
  • GameBench 2 Overall does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

DuelLab's overall GameBench 2 score: models write game-playing programs, the programs play each other, and the rating comes from match results. Shown only for models DuelLab ranks overall.

Published by
DuelLab GameBench 2
Category
Coding
Index weight
Reference only
Models
48
Data as of
2 Nov 2026
Newest result
26 Oct 2026

DuelLab GameBench 2, release gb2-20261026-cb69502a12d9 (https://benchmarks.duellab.org/releases/gb2-20261026-cb69502a12d9/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

Top 15
  1. 1GPT-6 Astra89.7
  2. 2Claude Opus 5.588.8
  3. 3Claude Fable 5.188.6
  4. 4GPT-6.1 Sol86.0
  5. 5Gemini 4 Argon83.3
  6. 6Claude Opus 578.1
  7. 7GPT-6 Sol74.7
  8. 8GPT-5.6 Sol68.7
  9. 9Claude Fable 567.6
  10. 10Muse Wonder66.4
  11. 11GLM 5.365.7
  12. 12Grok 4.765.0
  13. 13DeepSeek V4.1 Flash 101763.4
  14. 14Claude Sonnet 5.561.7
  15. 15Grok 4.661.0
48 of 48
#
1GPT-6 Astra89.768.22026-10-26
2Claude Opus 5.588.863.32026-10-26
3Claude Fable 5.188.666.52026-10-26
4GPT-6.1 Sol86.062.62026-10-26
5Gemini 4 Argon83.363.12026-10-24
6Claude Opus 578.165.32026-10-26
7GPT-6 Sol74.7–2026-10-26
8GPT-5.6 Sol68.755.12026-10-26
9Claude Fable 567.666.62026-10-26
10Muse Wonder66.464.02026-10-24
11GLM 5.365.7–2026-10-26
12Grok 4.765.0–2026-10-26
13DeepSeek V4.1 Flash 101763.4–2026-10-24
14Claude Sonnet 5.561.7–2026-10-26
15Grok 4.661.057.12026-10-26
16Gemini 3.8 Flash61.054.42026-10-26
17Muse Spark 1.359.757.82026-10-26
18Grok 4.558.359.82026-10-26
19GLM 5.3 FlashX57.6–2026-10-26
20GLM 5.3 Flash56.956.82026-10-26
21Qwen3.8 Max55.3–2026-10-26
22GPT-5.455.258.02026-10-26
23Qwen3.8 2.4T A95B54.652.02026-10-26
24DeepSeek V4.1 Flash53.7–2026-10-26
25Muse Spark 1.253.7–2026-10-26
26Gemini 3.6 Flash52.549.52026-10-26
27Ember-152.4–2026-10-26
28Claude Opus 4.851.660.02026-10-26
29Gemini 3.5 Flash51.148.72026-10-26
30GPT-5.550.761.72026-10-26
31GPT-6 Luna49–2026-10-26
32Claude Opus 4.548.957.92026-10-26
33Kimi K348.859.12026-10-26
34Claude Sonnet 548.155.92026-10-26
35GPT-5.6 Terra47.251.52026-10-26
36Beam 1.0 501B A23B44.754.52026-10-24
37GLM 5.244.551.72026-10-26
38DeepSeek V4 Pro43.853.22026-10-26
39GPT-5.6 Luna41.647.62026-10-26
40Seed-2.0-Code36.4–2026-10-26
41o336.261.02026-10-26
42DeepSeek V4 Flash34.254.32026-10-26
43Gemini 3.5 Flash Lite33.151.52026-10-26
44Inkling Small31.852.22026-10-26
45Muse Glimmer30.354.52026-10-26
46gpt-oss-120b30.245.52026-10-26
47Inkling26.747.52026-10-26
48Step 3.7 Flash24.253.02026-10-24

More coding benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on GameBench 2 Overall

Cite as: BenchLeader, “GameBench 2 Overall leaderboard”, data as of 2 Nov 2026, https://www.benchleader.com/benchmarks/duellab_gb2_overall.

GameBench 2 Overall: questions

What does GameBench 2 Overall measure?
DuelLab's overall GameBench 2 score: models write game-playing programs, the programs play each other, and the rating comes from match results. Shown only for models DuelLab ranks overall. Scores are reported in index; higher is better.
What is the top score on GameBench 2 Overall?
As of 2 Nov 2026, the highest score on GameBench 2 Overall is 89.7 by GPT-6 Astra (result dated 26 Oct 2026), ahead of Claude Opus 5.5 at 88.8. Scores come from DuelLab GameBench 2 and are re-read daily.
Which AI model leads GameBench 2 Overall?
GPT-6 Astra leads GameBench 2 Overall with 89.7 as of 2 Nov 2026 (result dated 26 Oct 2026), ahead of Claude Opus 5.5 at 88.8.
How many models have GameBench 2 Overall results?
48 model configurations have a GameBench 2 Overall result on BenchLeader, all taken from DuelLab GameBench 2.
Who runs GameBench 2 Overall and how often is it updated?
GameBench 2 Overall is published by DuelLab GameBench 2. BenchLeader re-reads the published results once a day and records the date each result was published.
Does GameBench 2 Overall count toward the BenchLeader Index?
No. GameBench 2 Overall is shown for reference but left out of the composite index.