Benchmark Almanac

LongProc

Gemini 1.5 Pro May 2024 74.2 leads since January 2025
Not in model cards

Models carry out long procedures such as converting HTML tables, tracing paths and tracking beliefs. The evaluator checks outputs as the required procedure grows.

Made by
Xi Ye and colleagues
Released
January 2025
Kind
Reasoning over long documents
Level
University
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code, Tables → Text, Tables
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude 3.5 Sonnet (October 2024) #150 on the capability index
Last model added
October 2024
Scores refreshed
November 20th, 2026
17
Entries on its own leaderboard, 14 models we track

Gemini 1.5 Pro leads with 74.2 among models released by October 2024 Gemini 1.5 Pro leads with 74.2 among models released by October 2024

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. The leaderboard reports a separate average at each length, so the headline is the plain mean of its three length averages. 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. This view shows only models we track, each at its best.

The newest model with a score here came out in October 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Gemini 1.5 Pro May 2024 74.2
  2. 2 GPT-4o Aug 2024 72.1
  3. 3 Claude 3.5 Sonnet (October 2024) 52.6
  4. 4 Llama 3.1-70B 51.7
  5. 5 Gemini 1.5 Flash May 2024 48.8
  6. 6 Qwen2.5-72B 45.0
  7. 7 GPT-4o mini 33.8
  8. 8 Qwen2.5-7B 18.2
  9. 9 Llama 3.1-8B 17.6
  10. 10 Phi-3-medium-128k-instruct 13.4
  11. 11 Qwen2.5-3B 11.4
  12. 12 Llama-3-8B-ProLong-512k-Inst 11.2
  13. 13 Mistral-7B-Inst-v0.3 10.7
  14. 14 Phi-3-small-128k-Inst 10.7
  15. 15 Jamba 1.5 Mini 9.70
  16. 16 Llama 3.2 3B 6.03
  17. 17 Llama 3.2 1B 1.37

Source: Official LongProc leaderboard

#EntryScoreSettingDateSource
1 Gemini 1.5 Pro May 2024 74.2 May 2024 Official leaderboard
2 GPT-4o Aug 2024 72.1 Aug 2024 Official leaderboard
3 Claude 3.5 Sonnet (October 2024) 52.6 Oct 2024 Official leaderboard
4 Llama 3.1-70B 51.7 Jul 2024 Official leaderboard
5 Gemini 1.5 Flash May 2024 48.8 May 2024 Official leaderboard
6 Qwen2.5-72B 45.0 Sep 2024 Official leaderboard
7 GPT-4o mini 33.8 Jul 2024 Official leaderboard
8 Qwen2.5-7B 18.2 Sep 2024 Official leaderboard
9 Llama 3.1-8B 17.6 Jul 2024 Official leaderboard
10 Phi-3-medium-128k-instruct 13.4 May 2024 Official leaderboard
11 Qwen2.5-3B 11.4 Sep 2024 Official leaderboard
12 Llama-3-8B-ProLong-512k-Inst 11.2 Official leaderboard
13 Mistral-7B-Inst-v0.3 10.7 Official leaderboard
14 Phi-3-small-128k-Inst 10.7 Official leaderboard
15 Jamba 1.5 Mini 9.70 Aug 2024 Official leaderboard
16 Llama 3.2 3B 6.03 Sep 2024 Official leaderboard
17 Llama 3.2 1B 1.37 Sep 2024 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 1.5 Pro May 2024 74.2 May 2024 Official leaderboard
2 GPT-4o Aug 2024 72.1 Aug 2024 Official leaderboard
3 Claude 3.5 Sonnet (October 2024) 52.6 Oct 2024 Official leaderboard
4 Llama 3.1-70B 51.7 Jul 2024 Official leaderboard
5 Gemini 1.5 Flash May 2024 48.8 May 2024 Official leaderboard
6 Qwen2.5-72B 45.0 Sep 2024 Official leaderboard
7 GPT-4o mini 33.8 Jul 2024 Official leaderboard
8 Qwen2.5-7B 18.2 Sep 2024 Official leaderboard
9 Llama 3.1-8B 17.6 Jul 2024 Official leaderboard
10 Phi-3-medium-128k-instruct 13.4 May 2024 Official leaderboard
11 Qwen2.5-3B 11.4 Sep 2024 Official leaderboard
12 Jamba 1.5 Mini 9.70 Aug 2024 Official leaderboard
13 Llama 3.2 3B 6.03 Sep 2024 Official leaderboard
14 Llama 3.2 1B 1.37 Sep 2024 Official leaderboard

Rankings by part

We read 19 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Average: 0.5K, Average: 2K and Average: 8K. GPT-4o leads 12 of them, and Gemini 1.5 Pro leads the other 7. Arrows show how far a model moved from its overall rank among the models we track.

Average: 0.5K

14 models we track have a score here.

  1. 1 GPT-4o Aug 2024 ↑1 94.8
  2. 2 Gemini 1.5 Pro May 2024 ↓1 89.2
  3. 3 Gemini 1.5 Flash May 2024 ↑2 78.9
  4. 4 Claude 3.5 Sonnet (October 2024) ↓1 78.4
  5. 5 Llama 3.1-70B ↓1 72.9
  6. 6 Qwen2.5-72B 68.7
  7. 7 GPT-4o mini 55.7
  8. 8 Qwen2.5-7B 27.8
  9. 9 Qwen2.5-3B ↑2 27.3
  10. 10 Llama 3.1-8B ↓1 26.9

Number one over time

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

Google DeepMind
Own leaderboard Not tested since Jan 2025

Source: Official LongProc 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
0.00 20.0 40.0 60.0 80.0 2024 2025 0.00 20.0 40.0 60.0 80.0 2024 2025

Source: Official LongProc 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 Long context

All 115 in Long context
  • MRCR v2 In use Multi-round coreference: retrieve the n-th of several similar requests hidden in a long synthetic conversation, up to 1M tokens. Muse Spark 1.3 98.5%
  • LongBench v2 In use 503 multiple-choice questions over 8K-2M-word contexts (QA, dialogue, code repos, structured data). Accuracy. Inkling 2 Small 70.6%
  • GraphWalks In use A graph is given as an edge list in long context. The model answers BFS or parent queries, scored by F1. Claude Fable 5 99.96%
  • OpenAI MRCR v2 In use Models read a long synthetic chat that hides 2, 4 or 8 identical writing requests and must reproduce the response to the i-th one. Scored by mean string match ratio per context bin. Gemini 3.7 Flash 89%
  • CL-bench In use Model must learn new knowledge from a supplied long context (rules, procedures, data) and solve tasks. Strict LLM-verified rubrics. GPT-5.4 27.9%
  • AA-LCR 1.1 In use Models answer 100 questions that need reasoning across sets of real documents of about 100K tokens. An LLM equality checker with a fixed system prompt compares each answer with the key. Step 5 Preview 88.3%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in October 2024, for Claude 3.5 Sonnet (October 2024). Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (17 entries).
  • Pass Models scored 14 models we track, plus 3 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in October 2024, for Claude 3.5 Sonnet (October 2024).
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude 3.5 Sonnet (October 2024) (#150).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on princeton-pli.github.io: 14 rows for models we track and 3 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