Benchmark Almanac

EmailBench

Claude Sonnet 4.5 33.5% leads since September 2026
Not in model cards

Agents handle email and productivity requests in a synthetic inbox. The evaluator combines executable checks and model-graded rubrics across 206 scenarios.

Made by
Microsoft
Released
September 2026
Kind
Documents and spreadsheets
Level
University
Scored
Graded by a model
Answers
Agent task
Measures
Capability
Language
English
In and out
Text → Actions, Text
Interaction
Agentic
Environment
APIs and tools
Items from
Generated
Refresh
One fixed set
Run by
A host
In views
Agents
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
7
Models scored in its paper

Claude Sonnet 4.5 leads with 33.5% among models released by April 2026

Best score per model from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best.

The newest model with a score here came out in April 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 Claude Sonnet 4.5 33.5%
  2. 2 Claude Opus 4.1 29.6%
  3. 3 GPT-5.4 Reasoning 28.2%
  4. 4 GPT-5.5 Reasoning 27.7%
  5. 5 GPT-5.2 22.8%
  6. 6 o3 22.3%
  7. 7 GPT-5 20.9%

Source: The EmailBench paper

#ModelScore ±SettingReleasedRun onSource
1 Claude Sonnet 4.5 33.5% 27.4 to 40.2 Sep 2025 May 2026 Official leaderboard
2 Claude Opus 4.1 29.6% 23.8 to 36.2 Aug 2025 May 2026 Official leaderboard
3 GPT-5.4 28.2% 22.5 to 34.7 Reasoning Mar 2026 May 2026 Official leaderboard
4 GPT-5.5 27.7% 22 to 34.1 Reasoning Apr 2026 May 2026 Official leaderboard
5 GPT-5.2 22.8% 17.6 to 29 Dec 2025 May 2026 Official leaderboard
6 o3 22.3% 17.2 to 28.5 Apr 2025 May 2026 Official leaderboard
7 GPT-5 20.9% 15.9 to 26.9 Aug 2025 May 2026 Official leaderboard

Rankings by part

We read 19 parts from the benchmark’s paper, each ranking models on one part of the test, such as filters, folders and settings. Claude Opus 4.1 leads 6 of them, and 4 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

filters

7 models we track have a score here.

  1. 1 Claude Opus 4.1 ↑1 90%
  2. 2 Claude Sonnet 4.5 ↓1 80%
  3. 3 o3 ↑3 50%
  4. 4 GPT-5.2 ↑1 50%
  5. 5 GPT-5.4 ↓2 40%
  6. 6 GPT-5.5 ↓2 40%
  7. 7 GPT-5 20%

Number one over time

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

Anthropic
Benchmark paper Not tested since Sep 2026

Source: The EmailBench 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 Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Sonnet 4.5 o3 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Sonnet 4.5 o3

Source: Official EmailBench 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 10 checks pass, 2 warnings. 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 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (7 entries).
  • Pass Models scored 7 models we track.
  • Warning Newest score The newest came 7 months ago, in April 2026, for GPT-5.5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
  • Pass Headroom The best is 33.5%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 8 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