Benchmark Almanac

AutoVend

GPT-6.1 Sol 309.4 leads since October 2026
Not in model cards

Agents buy stock and schedule deliveries for 20 simulated vending machines over 180 days. They share one warehouse and pay for each site visit.

StepFun reports five models with three runs each in its AutoVend technical report. The main score measures final net assets in thousands of yuan.

Made by
StepFun and collaborators
Released
October 2026
Kind
Running a business
Level
University
Scored
Checked automatically
Answers
Agent task
Measures
Capability
In and out
Text, Tables → Actions
Interaction
Agentic
Environment
Browser, A simulator
Items from
Generated
Refresh
One fixed set
Run by
A host
Strongest model tested
GPT-6.1 Sol #6 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
5
Models scored in its paper

GPT-6.1 Sol leads with 309.4 among models released by September 2026

Best score per model from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The score is mean final net assets, in thousands of yuan, over three runs. Where a model ran with several settings or harnesses, we keep its best.

The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-6.1 Sol 309.4
  2. 2 GLM-5.3-Flash 148.4
  3. 3 DeepSeek V4.1 Flash 133.9
  4. 4 Claude Sonnet 5.5 103.0
  5. 5 Kimi K3 91.5

Source: The AutoVend paper

#ModelScoreSettingDateSource
1 GPT-6.1 Sol 309.4 Sep 2026 Official leaderboard
2 GLM-5.3-Flash 148.4 Aug 2026 Official leaderboard
3 DeepSeek V4.1 Flash 133.9 Sep 2026 Official leaderboard
4 Claude Sonnet 5.5 103.0 Sep 2026 Official leaderboard
5 Kimi K3 91.5 Jul 2026 Official leaderboard

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI
Benchmark paper Not tested since Oct 2026
GPT-6.1 Sol 309.4

Source: The AutoVend paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper
100.0 200.0 300.0 100.0 200.0 300.0

Source: Official AutoVend leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Economic work

All 210 in Economic work
  • JobBench In use 130 tasks from 35 occupations workers most want delegated. Agent works through messy reference files, graded by fact-anchored binary rubrics. GLM-5.3 61.4%
  • GDPval In use Real work tasks from 44 occupations with reference files. Experts blindly compare model deliverables with expert work. Metric is win-or-tie rate. Seed2.1 Pro 87.9%
  • GDPval-AA v2.1 In use Agents with shell and web access complete 220 GDPval gold tasks from 44 occupations. Blind pairwise comparisons give Elo ratings. Claude Opus 5.5 1,846
  • AA-Briefcase v1.1 In use Version 1.1 of Artificial Analysis's multi-hour office work evaluation, with Elo ratings fitted by Crowd-BT. Claude Opus 5.5 1,822
  • Agents' Last Exam v1.1 In use Agents complete 122 public long-horizon professional workflow tasks across 55 sub-industries. Deterministic graders check outcomes against hidden references. GPT-6.2 Astra 47.5%
  • CoWorkBench In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished). Claude Fable 5 75.9%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the benchmark’s paper (5 entries).
  • Pass Models scored 5 models we track.
  • Pass Newest score The newest is from September 2026, for GPT-6.1 Sol.
  • Note Frontier models Only 2 of the top 10 have a score. The strongest tested is GPT-6.1 Sol (#6).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 5 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check