Benchmark Almanac

BLUR

HuggingFace Agents + Claude 3.5 Sonnet 56% tops its leaderboard OpenAI Computer-Using Agent (CUA) 54% leads as a model on its own
Not in model cards

Models identify a remembered item from incomplete clues in several languages and media.

The paper reports model results.

Made by
Patronus AI
Released
March 2025
Kind
Questions that need search
Level
Everyday
Scored
Checked automatically
Answers
Short answer
Tasks
Public, plus a private test set
Measures
Capability
In and out
Text, Images, Audio, Video → Actions, Text
Interaction
Agentic
Environment
Browser
Items from
Real work and use
Refresh
One fixed set
Run by
A host
Strongest model tested
o1 #118 on the capability index
Last model added
May 2025
Scores refreshed
November 20th, 2026
12
Entries on its own leaderboard, 6 models we track

HuggingFace Agents + Claude 3.5 Sonnet tops the leaderboard with 56% OpenAI Computer-Using Agent (CUA) leads with 54% among models released by May 2025

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. The best model on its own is OpenAI Computer-Using Agent (CUA), at 54%. Agents built around a model, such as HuggingFace Agents + Claude 3.5 Sonnet, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

The newest model with a score here came out in May 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 HuggingFace Agents + Claude 3.5 Sonnet 56%
  2. 2 OpenAI Computer-Using Agent (CUA) 54%
  3. 3 DynaSaur + GPT-4o Mar 2025 50%
  4. 4 ChatGPT-4o 49%
  5. 5 o1 49%
  6. 6 DeepSeek-R1 May 2025 41%
  7. 7 Claude 3.5 Sonnet 40%
  8. 8 GPT-4o Nov 2024 38%
  9. 9 Mistral Le Chat 37%
  10. 10 Llama 3.1-405B 30%
  11. 11 Microsoft Copilot 27%
  12. 12 Perplexity Pro Search 27%

Source: Official BLUR leaderboard

#EntryScoreSettingDateSource
1 HuggingFace Agents + Claude 3.5 Sonnet 56% Jun 2024 Official leaderboard
2 OpenAI Computer-Using Agent (CUA) 54% Jan 2025 Official leaderboard
3 DynaSaur + GPT-4o Mar 2025 50% Mar 2025 Official leaderboard
4 ChatGPT-4o 49% Official leaderboard
5 o1 49% Dec 2024 Official leaderboard
6 DeepSeek-R1 May 2025 41% May 2025 Official leaderboard
7 Claude 3.5 Sonnet 40% Jun 2024 Official leaderboard
8 GPT-4o Nov 2024 38% Nov 2024 Official leaderboard
9 Mistral Le Chat 37% Official leaderboard
10 Llama 3.1-405B 30% Jul 2024 Official leaderboard
11 Microsoft Copilot 27% Official leaderboard
12 Perplexity Pro Search 27% Official leaderboard
#ModelScoreSettingDateSource
1 OpenAI Computer-Using Agent (CUA) 54% Jan 2025 Official leaderboard
2 o1 49% Dec 2024 Official leaderboard
3 DeepSeek-R1 May 2025 41% May 2025 Official leaderboard
4 Claude 3.5 Sonnet 40% Jun 2024 Official leaderboard
5 GPT-4o Nov 2024 38% Nov 2024 Official leaderboard
6 Llama 3.1-405B 30% Jul 2024 Official leaderboard

Rankings by part

We read 5 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Text-only queries, Queries with file inputs and Easy. OpenAI Computer-Using Agent (CUA) leads 4 of them, and o1 leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Text-only queries

6 models we track have a score here.

  1. 1 OpenAI Computer-Using Agent (CUA) 57%
  2. 2 o1 54%
  3. 3 DeepSeek-R1 May 2025 45%
  4. 4 Claude 3.5 Sonnet 44%
  5. 5 GPT-4o Nov 2024 42%
  6. 6 Llama 3.1-405B 34%

Number one over time

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

OpenAI
Own leaderboard Not tested since May 2025

Source: Official BLUR leaderboard

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
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) OpenAI Computer-Using Agent (CUA) o1 Claude 3.5 Sonnet 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) OpenAI Computer-Using Agent (CUA) Claude 3.5 Sonnet

Source: Official BLUR 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 Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy. Qwen 3.6 Max (Preview) 73.9%
  • WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1. Tencent Hy4 preview 83.9%
  • BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 18 months ago, in May 2025, for DeepSeek-R1. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (12 entries).
  • Pass Models scored 6 models we track, plus 6 other entries on its leaderboard.
  • Warning Newest score The newest came 18 months ago, in May 2025, for DeepSeek-R1.
  • Note Frontier models None of the top 10 has a score. The strongest tested is o1 (#118).
  • Pass Headroom The best is 54%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on huggingface.co: 8 rows for models we track and 4 other entries 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