Benchmark Almanac

Remote Labor Index

Claude Opus 5.5 21.3% leads since Sep 2026 Not in model cards

Agents complete real paid remote freelance projects from many sectors. Reviewers judge whether each deliverable would be accepted as commissioned work, giving an automation rate.

The best agent automated only 2.5% of projects at launch. Only a 10 task public set is released.

In the authors’ words

AIs have made rapid progress on research-oriented benchmarks of knowledge and reasoning, but it remains unclear how these gains translate into economic value and automation. To measure this, we introduce the Remote Labor Index (RLI), a broadly multi-sector benchmark comprising real-world, economically valuable projects designed to evaluate end-to-end agent performance in practical settings. AI agents perform near the floor on RLI, with the highest-performing agent achieving an automation rate of 2.5%. These results help ground discussions of AI automation in empirical evidence, setting a common basis for tracking AI impacts and enabling stakeholders to proactively navigate AI-driven labor automation.

Abstract of Remote Labor Index: Measuring AI Automation of Remote Work (arXiv:2510.26787, CC0)
Made by Center for AI Safety and Scale AI Released Oct 2025 Tasks from real jobs Public tasks plus a private test set Last model added Sep 2026 Scores refreshed 2 Nov 2026
Level
Expert Specialists with years of training, often over hours of work
Scored
Graded by people Human experts grade or compare the outputs.
Answers
Open-ended Free-form text, documents, designs or media judged as a whole.
centerforaisafety/rli_evaluation_platform 82 stars Paper Dataset Leaderboard remotelabor.ai
20
Entries on its own leaderboard, 17 models we track

Claude Opus 5.5 leads with 21.3%

Best score per entry from the benchmark's own leaderboard, read on 2 Nov 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders.

20 entries. Scroll sideways for the rest.

  1. 21.3 Claude Opus 5.5
  2. 20.8 GPT-6 Astra
  3. 17.9 Claude Fable 5.1
  4. 17.1 Gemini 4 Argon
  5. 15.8 Claude Fable 5
  6. 8.33 Claude Opus 4.8
  7. 6.25 GPT-5.5
  8. 5.83 Gemini 3.8 Flash
  9. 5.00 Gemini 3.7 Flash
  10. 4.17 Claude Opus 4.6
  11. 3.75 Claude Opus 4.5
  12. 2.92 Manus_1.6
  13. 2.50 GPT-5.2
  14. 2.50 Manus 1.5
  15. 2.50 Manus 1.0
  16. 2.08 Claude Sonnet 4.5
  17. 1.67 GPT-5
  18. 1.25 ChatGPT agent
  19. 1.25 Gemini 3 Pro
  20. 0.83 Gemini 2.5 Pro
  1. 1 Claude Opus 5.5 21.3%
  2. 2 GPT-6 Astra 20.8%
  3. 3 Claude Fable 5.1 17.9%
  4. 4 Gemini 4 Argon 17.1%
  5. 5 Claude Fable 5 15.8%
  6. 6 Claude Opus 4.8 8.33%
  7. 7 GPT-5.5 6.25%
  8. 8 Gemini 3.8 Flash 5.83%
  9. 9 Gemini 3.7 Flash 5.00%
  10. 10 Claude Opus 4.6 4.17%
  11. 11 Claude Opus 4.5 thinking 3.75%
  12. 12 Manus_1.6 max 2.92%
  13. 13 GPT-5.2 medium 2.50%
  14. 14 Manus 1.5 2.50%
  15. 15 Manus 1.0 2.50%
  16. 16 Claude Sonnet 4.5 2.08%
  17. 17 GPT-5 1.67%
  18. 18 ChatGPT agent 1.25%
  19. 19 Gemini 3 Pro 1.25%
  20. 20 Gemini 2.5 Pro Jun 2025 0.83%
#EntryScoreSettingDateSource
1 Claude Opus 5.5 21.3% Sep 2026 Official leaderboard
2 GPT-6 Astra 20.8% Sep 2026 Official leaderboard
3 Claude Fable 5.1 17.9% Sep 2026 Official leaderboard
4 Gemini 4 Argon 17.1% Sep 2026 Official leaderboard
5 Claude Fable 5 15.8% Jun 2026 Official leaderboard
6 Claude Opus 4.8 8.33% May 2026 Official leaderboard
7 GPT-5.5 6.25% Codex Apr 2026 Official leaderboard
8 Gemini 3.8 Flash 5.83% Sep 2026 Official leaderboard
9 Gemini 3.7 Flash 5.00% Aug 2026 Official leaderboard
10 Claude Opus 4.6 4.17% Cowork Feb 2026 Official leaderboard
11 Claude Opus 4.5 3.75% thinking Nov 2025 Official leaderboard
12 Manus_1.6 2.92% max Official leaderboard
13 GPT-5.2 2.50% medium Dec 2025 Official leaderboard
14 Manus 1.5 2.50% Official leaderboard
15 Manus 1.0 2.50% Official leaderboard
16 Claude Sonnet 4.5 2.08% Sep 2025 Official leaderboard
17 GPT-5 1.67% Aug 2025 Official leaderboard
18 ChatGPT agent 1.25% Jul 2025 Official leaderboard
19 Gemini 3 Pro 1.25% Nov 2025 Official leaderboard
20 Gemini 2.5 Pro Jun 2025 0.83% Jun 2025 Official leaderboard

Epoch AI has scores for 16 models. We plot every source in the charts below.

9 models have held number one since its release in Oct 2025

Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and Epoch AI’s runs.

Anthropic OpenAI
Own leaderboard 9 leaders
Claude Opus 5.5 21.3%
Epoch AI runs 9 leaders
Claude Opus 5.5 21.3%

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 Run or collected by Epoch AI Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Opus 5.5 Claude Fable 5 Claude Opus 4.8 Claude Opus 4.6 Claude Sonnet 4.5 Gemini 2.5 Pro 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Opus 5.5 Claude Opus 4.8

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%