Benchmark Almanac

FoodTruck Bench

Claude Opus 5 $75,264 leads since Jul 2026 Not in model cards

Agents run a food truck in Austin, Texas for a 30-day simulation, choosing locations, menus, prices, inventory, staff and loans with 34 tools. Humans can play the same game.

60 models tested on six locations. Scored by business results at the end of the run.

Made by FoodTruck Bench maintainers Simulated companies All tasks public Last model added Oct 2026 Scores refreshed 2 Nov 2026
Level
Everyday Most adults, without training
Scored
Settled by events Predictions are scored against what later happens.
Answers
Agent task Many steps in a terminal, browser, desktop, app, game or robot.
Leaderboard foodtruckbench.com
10
Models scored on its own leaderboard

Claude Opus 5 leads with $75,264

Best score per model from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best.

  1. $75,264 Claude Opus 5
  2. $63,464 MiniMax-M3.1-Flash
  3. $62,732 Claude Fable 5.1
  4. $61,408 GPT-5.5
  5. $59,537 Grok 4.6
  6. $57,761 GPT-6 Astra
  7. $53,289 DeepSeek V4.1 Flash 1017
  8. $53,229 GPT-5.6 Sol
  9. $50,179 DeepSeek V4.1 Flash
  10. $49,519 Claude Opus 4.6
  1. 1 Claude Opus 5 $75,264
  2. 2 MiniMax-M3.1-Flash $63,464
  3. 3 Claude Fable 5.1 $62,732
  4. 4 GPT-5.5 $61,408
  5. 5 Grok 4.6 $59,537
  6. 6 GPT-6 Astra $57,761
  7. 7 DeepSeek V4.1 Flash 1017 $53,289
  8. 8 GPT-5.6 Sol $53,229
  9. 9 DeepSeek V4.1 Flash $50,179
  10. 10 Claude Opus 4.6 $49,519
#ModelScoreSettingDateSource
1 Claude Opus 5 $75,264 Jul 2026 Official leaderboard
2 MiniMax-M3.1-Flash $63,464 Oct 2026 Official leaderboard
3 Claude Fable 5.1 $62,732 Sep 2026 Official leaderboard
4 GPT-5.5 $61,408 Apr 2026 Official leaderboard
5 Grok 4.6 $59,537 Aug 2026 Official leaderboard
6 GPT-6 Astra $57,761 Sep 2026 Official leaderboard
7 DeepSeek V4.1 Flash 1017 $53,289 Oct 2026 Official leaderboard
8 GPT-5.6 Sol $53,229 Jul 2026 Official leaderboard
9 DeepSeek V4.1 Flash $50,179 Sep 2026 Official leaderboard
10 Claude Opus 4.6 $49,519 Feb 2026 Official leaderboard

3 models have held number one since 2023

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

Anthropic OpenAI
Own leaderboard 3 leaders
Claude Opus 5 $75,264

Scores over time

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

On the benchmark’s own leaderboard
$40,000 $50,000 $60,000 $70,000 $80,000 2026 $40,000 $50,000 $60,000 $70,000 $80,000 2026

No lab model card reports it yet

We read the headline results of 479 launch posts from 51 labs, and none of them list this benchmark.

More in Economic work

All 91 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
  • CoWorkBench In use Agent completes long-horizon productivity tasks across computer science, finance, law and medicine. Qwen in-house scoring (method unpublished). Qwen3.8-Omni-Flash 75.3%
  • OfficeQA In use Grounded reasoning over a large corpus of U.S. Treasury Bulletin documents: locate tables and compute precise numeric answers. Claude Opus 4.7 86.3%