BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

Vending-Bench 2 leaderboard

As of 6 Nov 2026, the top score on Vending-Bench 2 is 16484.2 by Gemini 4 Carbon (Google), ahead of GPT-6 Astra at 15514.7, across 71 model configurations with a published result (Andon Labs, result dated 28 Oct 2026). The agent manages a vending business over a simulated year: ordering, pricing, and dealing with suppliers; scored on final balance. Vending-Bench 2 is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Gemini 4 Carbon holds the top Vending-Bench 2 score at 16484.2 (Andon Labs, 28 Oct 2026).
  • GPT-6 Astra is second on Vending-Bench 2 at 15514.7 (Andon Labs, 3 Sep 2026).
  • Vending-Bench 2: Running a simulated vending-machine business for a year (Andon Labs).
  • 71 model configurations have a published Vending-Bench 2 result on BenchLeader (BenchLeader, 6 Nov 2026).
  • Vending-Bench 2 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Running a simulated vending-machine business for a year.

Published by
Andon Labsdata via Epoch AI Benchmarking Hub
Category
Agents & tools
Index weight
Reference only
Models
71
Data as of
6 Nov 2026
Newest result
28 Oct 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 agent manages a vending business over a simulated year: ordering, pricing, and dealing with suppliers; scored on final balance.

How it is scored

Final balance in simulated dollars, as published by Andon Labs.

What to keep in mind

A dollar figure, not a percent; long-horizon coherence is what it tests.

Top 15
  1. 1Gemini 4 Carbon16484.2
  2. 2GPT-6 Astra15514.7
  3. 3GPT-6 Sol14427.9
  4. 4Gemini 4 Argon13718.2
  5. 5Claude Opus 511181.9
  6. 6Claude Sonnet 5.511165.3
  7. 7Claude Opus 4.710936.8
  8. 8Grok 4.710536.8
  9. 9GPT-5.6 Sol9619.4
  10. 10Muse Wonder9432.2
  11. 11Claude Opus 5.59235.3
  12. 12GPT-6.1 Terra9107.5
  13. 13Grok 4.69047.0
  14. 14GLM 5.28313.8
  15. 15GLM 5.38163.6
71 of 71
#
1Gemini 4 Carbon16484.2–2026-10-28
2GPT-6 Astra15514.768.22026-09-03
3GPT-6 Sol14427.9–2026-09-22
4Gemini 4 Argon13718.263.42026-09-30
5Claude Opus 511181.965.22026-07-24
6Claude Sonnet 5.511165.3–2026-09-28
7Claude Opus 4.710936.863.12026-04-16
8Grok 4.710536.8–2026-09-21
9GPT-5.6 Sol9619.455.02026-07-09
10Muse Wonder9432.263.92026-10-12
11Claude Opus 5.59235.363.42026-09-22
12GPT-6.1 Terra9107.5–2026-10-26
13Grok 4.69047.057.12026-08-12
14GLM 5.28313.851.52026-06-16
15GLM 5.38163.6–2026-08-14
16Claude Opus 4.68017.662.32026-02-05
17GPT-5.57523.861.72026-04-23
18GPT-5.6 Terra7343.251.52026-07-09
19Claude Sonnet 4.67204.158.22026-02-17
20Muse Spark 1.16520.562.22026-07-09
21Claude Sonnet 56377.755.92026-06-30
22Kimi K2.66204.655.22026-04-20
23GPT-5.46144.258.02026-03-05
24GPT-5.3-Codex5940.1–2026-02-05
25Claude Opus 4.85787.460.02026-05-28
26Claude Fable 5high5680.364.32026-06-09
27GLM 5.15634.455.72026-04-07
28Gemini 3 Pro5478.260.22025-11-18
29Claude Fable 5.15421.666.42026-09-01
30Gemini 3.5 Flash5396.448.62026-05-19
31Kimi K35165.059.12026-07-16
32Qwen3.6 Plus5114.959.12026-03-31
33Gemini 3.8 Flash5093.854.42026-09-02
34Kimi K2.7 Code5082.954.22026-06-12
35Claude Fable 5low5018.5–2026-06-09
36Claude Opus 4.54967.157.92025-11-24
37Claude Fable 5max4966.665.32026-06-09
38Grok 4.204662.9–2026-02-17
39Claude Fable 54529.966.52026-06-09
40GLM-54432.154.82026-02-11
41Claude Fable 5medium4339.8–2026-06-09
42Qwen3.6 Maxmax4254.262.42026-04-20
43GPT-5.6 Luna4094.747.62026-07-09
44Grok 4.53887.459.82026-07-08
45Claude Sonnet 4.53838.753.72025-09-29
46Gemini 3.1 Pro Customtools3774.3–2026-02-19
47Gemini 3 Flash3634.756.92025-12-17
48Claude Haiku 5.53602.5–2026-10-07
49GPT-5.23591.357.92025-12-11
50DeepSeek V4 Pro3284.553.22026-04-24
51Claude Opus 4.8max2992.360.22026-05-28
52GLM 4.72376.851.72025-12-22
53MiniMax M32157.856.42026-06-01
54GPT-5.11473.456.82025-11-13
55Kimi K2.51198.554.22026-01-27
56Grok 4.1 Fastthinking1106.655.62025-11-19
57DeepSeek V3.2103456.52025-09-29
58Gemini 3.1 Pro911.260.52026-02-19
59Gemini 2.5 Pro573.654.52025-06-17
60Gemini 2.5 Flash548.852.72025-06-17

More agents & tools benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on Vending-Bench 2

Cite as: BenchLeader, “Vending-Bench 2 leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vending_bench_2.

Vending-Bench 2: questions

What does Vending-Bench 2 measure?
The agent manages a vending business over a simulated year: ordering, pricing, and dealing with suppliers; scored on final balance. Scores are reported in score; higher is better.
What is the top score on Vending-Bench 2?
As of 6 Nov 2026, the highest score on Vending-Bench 2 is 16484.2 by Gemini 4 Carbon (result dated 28 Oct 2026), ahead of GPT-6 Astra at 15514.7. Scores come from Andon Labs and are re-read daily.
Which AI model leads Vending-Bench 2?
Gemini 4 Carbon leads Vending-Bench 2 with 16484.2 as of 6 Nov 2026 (result dated 28 Oct 2026), ahead of GPT-6 Astra at 15514.7.
How many models have Vending-Bench 2 results?
71 model configurations have a Vending-Bench 2 result on BenchLeader, all taken from Andon Labs via Epoch AI Benchmarking Hub.
Who runs Vending-Bench 2 and how often is it updated?
Vending-Bench 2 is published by Andon Labs. BenchLeader re-reads the published results once a day and records the date each result was published.
Does Vending-Bench 2 count toward the BenchLeader Index?
No. Vending-Bench 2 is shown for reference but left out of the composite index.