EKWbench
Agents complete operational tasks such as sourcing reviews and referral triage. The score averages the share of verifier weight earned.
The board distinguishes weighted reward from the all-core pass rate.
- Made by
- Turing
- Kind
- Tasks from real jobs
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- Kept private
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Terminal
- Refresh
- One fixed set
- Run by
- A host
- Occupations
- Management, Business and financial operations, Arts, design, entertainment, sports and media, Healthcare practitioners and technical
- In views
- Agents, Biology and medicine, Law and finance
- Strongest model tested
- Claude Opus 5 #11 on the capability index
- Last model added
- July 2026
- Scores refreshed
- November 20th, 2026
Claude Opus 5 leads with 74.4% among models released by July 2026
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in July 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 13 more
- 1 Claude Opus 5 max 74.4%
- 2 GPT-5.6 Sol max 72.5%
- 3 GPT-5.6 Terra max 68.3%
- 4 Claude Sonnet 5 max 66.9%
- 5
Gemini 3.6 Flash high 53.7%
- 6
Gemini 3.1 Pro high 37.4%
Source: Official EKWbench leaderboard
| # | Model | Score ± | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 5 | 74.4% ±2.7 | max, claude-code 2.1.237 | Jul 2026 | Official leaderboard |
| 2 | GPT-5.6 Sol | 72.5% ±2.7 | max, codex 0.148.0 | Jul 2026 | Official leaderboard |
| 3 | GPT-5.6 Terra | 68.3% ±2.7 | max, codex 0.148.0 | Jul 2026 | Official leaderboard |
| 4 | Claude Sonnet 5 | 66.9% ±3 | max, claude-code 2.1.237 | Jun 2026 | Official leaderboard |
| 5 | | 53.7% ±3.1 | high, gemini-cli 0.56.0 | Jul 2026 | Official leaderboard |
| 6 | | 37.4% ±2.8 | high, gemini-cli 0.56.0 | Feb 2026 | Official leaderboard |
Rankings by part
We read 8 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Finance and insurance, Retail trade and Manufacturing. Claude Opus 5 leads 6 of them, and GPT-5.6 Sol leads the other 2. Arrows show how far a model moved from its overall rank among the models we track.
Finance and insurance
6 models we track have a score here.
- 1 Claude Opus 5 72.9%
- 2 GPT-5.6 Sol 70.7%
- 3 GPT-5.6 Terra 65.5%
- 4 Claude Sonnet 5 61.9%
- 5
Gemini 3.6 Flash 49.2%
- 6
Gemini 3.1 Pro 34.5%
Retail trade
6 models we track have a score here.
- 1 GPT-5.6 Sol ↑1 72.7%
- 2 Claude Opus 5 ↓1 70.8%
- 3 GPT-5.6 Terra 66%
- 4 Claude Sonnet 5 64%
- 5
Gemini 3.6 Flash 53.3%
- 6
Gemini 3.1 Pro 31%
Manufacturing
6 models we track have a score here.
- 1 Claude Opus 5 73.2%
- 2 GPT-5.6 Sol 71.6%
- 3 GPT-5.6 Terra 69.2%
- 4 Claude Sonnet 5 63%
- 5
Gemini 3.6 Flash 48.5%
- 6
Gemini 3.1 Pro 36.1%
Health care and social assistance
6 models we track have a score here.
- 1 Claude Opus 5 73%
- 2 GPT-5.6 Sol 68.1%
- 3 Claude Sonnet 5 ↑1 67.7%
- 4 GPT-5.6 Terra ↓1 65.7%
- 5
Gemini 3.6 Flash 53.2%
- 6
Gemini 3.1 Pro 35.1%
Information
6 models we track have a score here.
- 1 Claude Opus 5 76.8%
- 2 GPT-5.6 Sol 75%
- 3 GPT-5.6 Terra 68.3%
- 4 Claude Sonnet 5 67.4%
- 5
Gemini 3.6 Flash 52.1%
- 6
Gemini 3.1 Pro 40.4%
Real estate and rental and leasing
6 models we track have a score here.
- 1 Claude Opus 5 83.8%
- 2 Claude Sonnet 5 ↑2 83.6%
- 3 GPT-5.6 Sol ↓1 75.7%
- 4 GPT-5.6 Terra ↓1 74.8%
- 5
Gemini 3.6 Flash 62%
- 6
Gemini 3.1 Pro 49.3%
Government
6 models we track have a score here.
- 1 Claude Opus 5 95.3%
- 2 GPT-5.6 Sol 80.3%
- 3 Claude Sonnet 5 ↑1 76.5%
- 4 GPT-5.6 Terra ↓1 76.5%
- 5
Gemini 3.6 Flash 62.7%
- 6
Gemini 3.1 Pro 41.8%
Professional, scientific, and technical services
6 models we track have a score here.
- 1 GPT-5.6 Sol ↑1 83.3%
- 2 GPT-5.6 Terra ↑1 82.5%
- 3 Claude Sonnet 5 ↑1 82.1%
- 4
Gemini 3.6 Flash ↑1 81.2%
- 5 Claude Opus 5 ↓4 78%
- 6
Gemini 3.1 Pro 59.3%
4 models have held number one on the official leaderboard since 2023
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official EKWbench leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official EKWbench 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
- 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
All 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access The tasks are private.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (6 entries).
- Pass Models scored 6 models we track.
- Pass Newest score The newest is from July 2026, for Claude Opus 5.
- Note Frontier models None of the top 10 has a score. The strongest tested is Claude Opus 5 (#11).
- Pass Headroom The best is 74.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on bench.turing.com: 6 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.