ParseBench-2
Version 2
Parsers extract text, formatting, tables, charts, layout, and forms from 4,317 page cases across 2,645 documents. Rule checks score six dimensions and average them equally.
Version 2 changes the tasks and scoring from ParseBench. Version 2.01 fixes annotations without defining another benchmark. Severity caps limit credit for serious errors.
- Made by
- LlamaIndex
- Released
- November 2026
- Kind
- Documents, charts and text in images
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- New versions replace it
- Run by
- Anyone
- Based on
- ParseBench
- Strongest model tested
- Claude Opus 5.5 #5 on the capability index
- Last model added
- November 2026
- Scores refreshed
- November 18th, 2026
LlamaParse (Agentic Plus) tops the leaderboard with 90.8% Claude Opus 5.5 leads with 87%
Best score per entry from the benchmark's own leaderboard, read on November 18th, 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. The best model we track is Claude Opus 5.5, at 87%. This view shows only models we track, each at its best.
- 90.8 LlamaParse (Agentic Plus)
- 87.0 Claude Opus 5.5
- 86.9 LlamaParse (Agentic)
- 86.1 Reducto (r-1)
- 85.4 LlamaParse (Cost Effective)
- 85.2
Gemini 4 Argon
- 84.8 Reducto
- 84.0 Extend (Parse 2)
- 83.9 Pulse (Ultra 2)
- 82.5 Datalab (Accurate)
- 82.4 Extend (Parse Light)
- 82.3 Datalab (Balanced)
- 80.1 KDL-Frontier-Parser-nano-v2
- 79.5 Datalab (Fast)
- 78.4 Infinity-Parser2 Pro
- 78.3 LightOnOCR-3 4B
- 78.1 GPT-5.6 Terra
- 77.7 Databricks AI Parse
- 77.6
Gemini 3 Flash
- 77.5 Chandra 2
- 77.1 LightOnOCR-3 0.8B
- 76.8 MinerU2.5-Pro
- 76.5
Gemini 3.7 Flash
- 76.5
Gemini 3.8 Flash
- 75.7 Infinity-Parser2 Flash
- 75.7
Gemini 3.5 Flash
- 75.7
Gemini 3.6 Flash
- 75.5
Muse Ember
- 75.2 GPT-6.1 Terra
- 74.9 Claude Sonnet 5.5
- 74.5 GPT-6 Luna
- 74.4
Qwen 3.8 27B
- 74.2 Nemotron Parse 3.5
- 73.9 Mistral OCR 4.1 (annotation)
- 72.5 Reducto (r-1-mini)
- 72.1 LandingAI
- 71.4 PaddleOCR-VL 1.6
- 71.1 Claude Haiku 5.5
- 71.0
MiniCPM-OCR-1.0
- 70.2 Extend (Parse)
- 70.1 GPT-5.6 Luna
- 69.2 Mistral OCR 4.1
- 68.9 surya-ocr-2
- 68.0 HunyuanOCR
- 67.2 Reducto (non-agentic)
- 66.7 Azure Document Intelligence (layout)
- 65.8 TeleOCR
- 65.5
Unlimited OCR 2
- 64.0 dots.mocr
- 63.5
Gemini 3.1 Flash-Lite
- 62.1
Gemini 3.5 Flash-Lite
- 57.8 Falcon-OCR
- 48.6 jina-ocr-v1
- 42.9
Qwen3.5-4B
- 42.7
Gemma 4 E4B
- 1 LlamaParse (Agentic Plus) ✓ 90.8%
- 2 Claude Opus 5.5 ✓ 87%
- 3 LlamaParse (Agentic) ✓ 86.9%
- 4 Reducto (r-1) ✓ 86.1%
- 5 LlamaParse (Cost Effective) ✓ 85.4%
- 6
Gemini 4 Argon low ✓ 85.2%
- 7 Reducto ✓ 84.8%
- 8 Extend (Parse 2) ✓ 84%
- 9 Pulse (Ultra 2) ✓ 83.9%
- 10 Datalab (Accurate) ✓ 82.5%
- 11 Extend (Parse Light) ✓ 82.4%
- 12 Datalab (Balanced) ✓ 82.3%
- 13 KDL-Frontier-Parser-nano-v2 ✓ 80.1%
- 14 Datalab (Fast) ✓ 79.5%
- 15 Infinity-Parser2 Pro ✓ 78.4%
- 16 LightOnOCR-3 4B ✓ 78.3%
- 17 GPT-5.6 Terra low ✓ 78.1%
- 18 Databricks AI Parse ✓ 77.7%
- 19
Gemini 3 Flash minimal ✓ 77.6%
- 20 Chandra 2 ✓ 77.5%
- 21 LightOnOCR-3 0.8B ✓ 77.1%
- 22 MinerU2.5-Pro ✓ 76.8%
- 23
Gemini 3.7 Flash low ✓ 76.5%
- 24
Gemini 3.8 Flash low ✓ 76.5%
- 25 Infinity-Parser2 Flash ✓ 75.7%
- 26
Gemini 3.5 Flash ✓ 75.7%
- 27
Gemini 3.6 Flash ✓ 75.7%
- 28
Muse Ember low ✓ 75.5%
- 29 GPT-6.1 Terra low ✓ 75.2%
- 30 Claude Sonnet 5.5 ✓ 74.9%
- 31 GPT-6 Luna low ✓ 74.5%
- 32
Qwen 3.8 27B thinking ✓ 74.4%
- 33 Nemotron Parse 3.5 ✓ 74.2%
- 34 Mistral OCR 4.1 (annotation) ✓ 73.9%
- 35 Reducto (r-1-mini) ✓ 72.5%
- 36 LandingAI ✓ 72.1%
- 37 PaddleOCR-VL 1.6 ✓ 71.4%
- 38 Claude Haiku 5.5 ✓ 71.1%
- 39
MiniCPM-OCR-1.0 ✓ 71%
- 40 Extend (Parse) ✓ 70.2%
- 41 GPT-5.6 Luna low ✓ 70.1%
- 42 Mistral OCR 4.1 ✓ 69.2%
- 43 surya-ocr-2 ✓ 68.9%
- 44 HunyuanOCR ✓ 68%
- 45 Reducto (non-agentic) ✓ 67.2%
- 46 Azure Document Intelligence (layout) ✓ 66.7%
- 47 TeleOCR ✓ 65.8%
- 48
Unlimited OCR 2 ✓ 65.5%
- 49 dots.mocr ✓ 64%
- 50
Gemini 3.1 Flash-Lite ✓ 63.5%
- 51
Gemini 3.5 Flash-Lite ✓ 62.1%
- 52 Falcon-OCR ✓ 57.8%
- 53 jina-ocr-v1 ✓ 48.6%
- 54
Qwen3.5-4B ✓ 42.9%
- 55
Gemma 4 E4B ✓ 42.7%
Source: Official ParseBench-2 leaderboard
- 87.0 Claude Opus 5.5
- 85.2
Gemini 4 Argon
- 78.1 GPT-5.6 Terra
- 77.6
Gemini 3 Flash
- 76.5
Gemini 3.7 Flash
- 76.5
Gemini 3.8 Flash
- 75.7
Gemini 3.5 Flash
- 75.7
Gemini 3.6 Flash
- 75.5
Muse Ember
- 75.2 GPT-6.1 Terra
- 74.9 Claude Sonnet 5.5
- 74.5 GPT-6 Luna
- 74.4
Qwen 3.8 27B
- 71.1 Claude Haiku 5.5
- 71.0
MiniCPM-OCR-1.0
- 70.1 GPT-5.6 Luna
- 65.5
Unlimited OCR 2
- 63.5
Gemini 3.1 Flash-Lite
- 62.1
Gemini 3.5 Flash-Lite
- 42.9
Qwen3.5-4B
- 42.7
Gemma 4 E4B
- 1 Claude Opus 5.5 ✓ 87%
- 2
Gemini 4 Argon low ✓ 85.2%
- 3 GPT-5.6 Terra low ✓ 78.1%
- 4
Gemini 3 Flash minimal ✓ 77.6%
- 5
Gemini 3.7 Flash low ✓ 76.5%
- 6
Gemini 3.8 Flash low ✓ 76.5%
- 7
Gemini 3.5 Flash ✓ 75.7%
- 8
Gemini 3.6 Flash ✓ 75.7%
- 9
Muse Ember low ✓ 75.5%
- 10 GPT-6.1 Terra low ✓ 75.2%
- 11 Claude Sonnet 5.5 ✓ 74.9%
- 12 GPT-6 Luna low ✓ 74.5%
- 13
Qwen 3.8 27B thinking ✓ 74.4%
- 14 Claude Haiku 5.5 ✓ 71.1%
- 15
MiniCPM-OCR-1.0 ✓ 71%
- 16 GPT-5.6 Luna low ✓ 70.1%
- 17
Unlimited OCR 2 ✓ 65.5%
- 18
Gemini 3.1 Flash-Lite ✓ 63.5%
- 19
Gemini 3.5 Flash-Lite ✓ 62.1%
- 20
Qwen3.5-4B ✓ 42.9%
- 21
Gemma 4 E4B ✓ 42.7%
Source: Official ParseBench-2 leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | LlamaParse (Agentic Plus) ✓ | 90.8% | Official leaderboard | ||
| 2 | Claude Opus 5.5 ✓ | 87% | Sep 2026 | Official leaderboard | |
| 3 | LlamaParse (Agentic) ✓ | 86.9% | Official leaderboard | ||
| 4 | Reducto (r-1) ✓ | 86.1% | Official leaderboard | ||
| 5 | LlamaParse (Cost Effective) ✓ | 85.4% | Official leaderboard | ||
| 6 | | 85.2% | low | Sep 2026 | Official leaderboard |
| 7 | Reducto ✓ | 84.8% | Official leaderboard | ||
| 8 | Extend (Parse 2) ✓ | 84% | Official leaderboard | ||
| 9 | Pulse (Ultra 2) ✓ | 83.9% | Official leaderboard | ||
| 10 | Datalab (Accurate) ✓ | 82.5% | Official leaderboard | ||
| 11 | Extend (Parse Light) ✓ | 82.4% | Official leaderboard | ||
| 12 | Datalab (Balanced) ✓ | 82.3% | Official leaderboard | ||
| 13 | KDL-Frontier-Parser-nano-v2 ✓ open weights | 80.1% | Official leaderboard | ||
| 14 | Datalab (Fast) ✓ | 79.5% | Official leaderboard | ||
| 15 | Infinity-Parser2 Pro ✓ open weights | 78.4% | Official leaderboard | ||
| 16 | LightOnOCR-3 4B ✓ open weights | 78.3% | Official leaderboard | ||
| 17 | GPT-5.6 Terra ✓ | 78.1% | low | Jul 2026 | Official leaderboard |
| 18 | Databricks AI Parse ✓ | 77.7% | Official leaderboard | ||
| 19 | | 77.6% | minimal | Dec 2025 | Official leaderboard |
| 20 | Chandra 2 ✓ open weights | 77.5% | Official leaderboard | ||
| 21 | LightOnOCR-3 0.8B ✓ open weights | 77.1% | Official leaderboard | ||
| 22 | MinerU2.5-Pro ✓ open weights | 76.8% | Official leaderboard | ||
| 23 | | 76.5% | low | Aug 2026 | Official leaderboard |
| 24 | | 76.5% | low | Sep 2026 | Official leaderboard |
| 25 | Infinity-Parser2 Flash ✓ open weights | 75.7% | Official leaderboard | ||
| 26 | | 75.7% | May 2026 | Official leaderboard | |
| 27 | | 75.7% | Jul 2026 | Official leaderboard | |
| 28 | | 75.5% | low | Nov 2026 | Official leaderboard |
| 29 | GPT-6.1 Terra ✓ | 75.2% | low | Oct 2026 | Official leaderboard |
| 30 | Claude Sonnet 5.5 ✓ | 74.9% | Sep 2026 | Official leaderboard | |
| 31 | GPT-6 Luna ✓ | 74.5% | low | Sep 2026 | Official leaderboard |
| 32 | | 74.4% | thinking | Aug 2026 | Official leaderboard |
| 33 | Nemotron Parse 3.5 ✓ open weights | 74.2% | Official leaderboard | ||
| 34 | Mistral OCR 4.1 (annotation) ✓ | 73.9% | Official leaderboard | ||
| 35 | Reducto (r-1-mini) ✓ | 72.5% | Official leaderboard | ||
| 36 | LandingAI ✓ | 72.1% | Official leaderboard | ||
| 37 | PaddleOCR-VL 1.6 ✓ open weights | 71.4% | Official leaderboard | ||
| 38 | Claude Haiku 5.5 ✓ | 71.1% | Oct 2026 | Official leaderboard | |
| 39 | | 71% | Oct 2026 | Official leaderboard | |
| 40 | Extend (Parse) ✓ | 70.2% | Official leaderboard | ||
| 41 | GPT-5.6 Luna ✓ | 70.1% | low | Jul 2026 | Official leaderboard |
| 42 | Mistral OCR 4.1 ✓ | 69.2% | Official leaderboard | ||
| 43 | surya-ocr-2 ✓ open weights | 68.9% | Official leaderboard | ||
| 44 | HunyuanOCR ✓ open weights | 68% | Official leaderboard | ||
| 45 | Reducto (non-agentic) ✓ | 67.2% | Official leaderboard | ||
| 46 | Azure Document Intelligence (layout) ✓ | 66.7% | Official leaderboard | ||
| 47 | TeleOCR ✓ open weights | 65.8% | Official leaderboard | ||
| 48 | | 65.5% | Oct 2026 | Official leaderboard | |
| 49 | dots.mocr ✓ open weights | 64% | Official leaderboard | ||
| 50 | | 63.5% | Mar 2026 | Official leaderboard | |
| 51 | | 62.1% | Jul 2026 | Official leaderboard | |
| 52 | Falcon-OCR ✓ open weights | 57.8% | Official leaderboard | ||
| 53 | jina-ocr-v1 ✓ open weights | 48.6% | Official leaderboard | ||
| 54 | | 42.9% | Feb 2026 | Official leaderboard | |
| 55 | | 42.7% | Mar 2026 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 ✓ | 87% | Sep 2026 | Official leaderboard | |
| 2 | | 85.2% | low | Sep 2026 | Official leaderboard |
| 3 | GPT-5.6 Terra ✓ | 78.1% | low | Jul 2026 | Official leaderboard |
| 4 | | 77.6% | minimal | Dec 2025 | Official leaderboard |
| 5 | | 76.5% | low | Aug 2026 | Official leaderboard |
| 6 | | 76.5% | low | Sep 2026 | Official leaderboard |
| 7 | | 75.7% | May 2026 | Official leaderboard | |
| 8 | | 75.7% | Jul 2026 | Official leaderboard | |
| 9 | | 75.5% | low | Nov 2026 | Official leaderboard |
| 10 | GPT-6.1 Terra ✓ | 75.2% | low | Oct 2026 | Official leaderboard |
| 11 | Claude Sonnet 5.5 ✓ | 74.9% | Sep 2026 | Official leaderboard | |
| 12 | GPT-6 Luna ✓ | 74.5% | low | Sep 2026 | Official leaderboard |
| 13 | | 74.4% | thinking | Aug 2026 | Official leaderboard |
| 14 | Claude Haiku 5.5 ✓ | 71.1% | Oct 2026 | Official leaderboard | |
| 15 | | 71% | Oct 2026 | Official leaderboard | |
| 16 | GPT-5.6 Luna ✓ | 70.1% | low | Jul 2026 | Official leaderboard |
| 17 | | 65.5% | Oct 2026 | Official leaderboard | |
| 18 | | 63.5% | Mar 2026 | Official leaderboard | |
| 19 | | 62.1% | Jul 2026 | Official leaderboard | |
| 20 | | 42.9% | Feb 2026 | Official leaderboard | |
| 21 | | 42.7% | Mar 2026 | Official leaderboard |
Rankings by part
We read 6 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Text, Formatting and Tables. Claude Opus 5.5 leads 3 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Text
21 models we track have a score here.
- 1 Claude Opus 5.5 92.6%
- 2 Claude Sonnet 5.5 ↑9 91.3%
- 3
Muse Ember ↑6 90.1%
- 4
Gemini 4 Argon ↓2 89.6%
- 5 Claude Haiku 5.5 ↑9 89.2%
- 6
Gemini 3.5 Flash ↑1 88.4%
- 7
Gemini 3.6 Flash ↑1 88.1%
- 8 GPT-6.1 Terra ↑2 87.3%
- 9
Gemini 3.8 Flash ↓3 87%
- 10
Gemini 3 Flash ↓6 86.7%
Formatting
21 models we track have a score here.
- 1 Claude Opus 5.5 89.4%
- 2
Gemini 4 Argon 89.2%
- 3 GPT-6 Luna ↑9 89.2%
- 4 GPT-5.6 Luna ↑12 88.5%
- 5 GPT-6.1 Terra ↑5 88.2%
- 6 GPT-5.6 Terra ↓3 87.9%
- 7 Claude Haiku 5.5 ↑7 87.5%
- 8
Gemini 3.8 Flash ↓2 86.7%
- 9
Gemini 3.7 Flash ↓4 86.6%
- 10 Claude Sonnet 5.5 ↑1 86.2%
Tables
21 models we track have a score here.
- 1 Claude Opus 5.5 94.2%
- 2
Gemini 4 Argon 92.8%
- 3 GPT-6.1 Terra ↑7 92.3%
- 4 Claude Sonnet 5.5 ↑7 91%
- 5
Gemini 3.5 Flash ↑2 89.9%
- 6
Gemini 3.6 Flash ↑2 89.8%
- 7
Gemini 3.7 Flash ↓2 89.5%
- 8
Muse Ember ↑1 89.1%
- 9
Gemini 3 Flash ↓5 89%
- 10
Gemini 3.8 Flash ↓4 88.9%
Charts
21 models we track have a score here.
- 1 GPT-5.6 Terra ↑2 76.3%
- 2 Claude Opus 5.5 ↓1 74.1%
- 3
Gemini 4 Argon ↓1 68.2%
- 4
Qwen 3.8 27B ↑9 65.2%
- 5
Muse Ember ↑4 60.5%
- 6
Gemini 3 Flash ↓2 58%
- 7 GPT-6 Luna ↑5 47.9%
- 8 GPT-6.1 Terra ↑2 46.7%
- 9 GPT-5.6 Luna ↑7 40.7%
- 10
MiniCPM-OCR-1.0 ↑5 40.2%
Visual grounding
21 models we track have a score here.
- 1
Gemini 4 Argon ↑1 89.8%
- 2 Claude Opus 5.5 ↓1 89.2%
- 3
Unlimited OCR 2 ↑14 85.4%
- 4
Gemini 3.8 Flash ↑2 82.8%
- 5
MiniCPM-OCR-1.0 ↑10 82.5%
- 6
Gemini 3.7 Flash ↓1 82.4%
- 7
Gemini 3.6 Flash ↑1 77.9%
- 8
Gemini 3.5 Flash ↓1 77.2%
- 9 Claude Haiku 5.5 ↑5 75.3%
- 10
Qwen 3.8 27B ↑3 70.5%
Forms
21 models we track have a score here.
- 1 GPT-6.1 Terra ↑9 83.5%
- 2 Claude Opus 5.5 ↓1 82.3%
- 3
Gemini 3.6 Flash ↑5 81.9%
- 4 Claude Sonnet 5.5 ↑7 81.7%
- 5
Muse Ember ↑4 81.7%
- 6
Gemini 4 Argon ↓4 81.4%
- 7
Gemini 3.5 Flash 81.2%
- 8
Gemini 3.8 Flash ↓2 80.5%
- 9
Gemini 3.7 Flash ↓4 80.2%
- 10 GPT-6 Luna ↑2 79.8%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official ParseBench-2 leaderboard
Claude Opus 5.5 on the official leaderboard passed 85% in September 2026
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official ParseBench-2 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.
Versions
We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.
Source: Official leaderboards
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy.
Qwen 3.6 Plus 86%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy.
Seed2.1 Pro 90.7%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7%
- MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy.
Qwen 3.8 Max 85.9%
Data audit
All 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (55 entries).
- Pass Models scored 21 models we track, plus 34 other entries on its leaderboard.
- Pass Newest score The newest is from November 2026, for Muse Ember.
- Pass Frontier models 4 of the top 10 have a score, led by Claude Opus 5.5 (#5).
- Note Headroom Saturated. The best is 87%, past 85% since September 2026.
Leaderboard import 3 pass
- Pass Importer We read the board on parsebench.ai: 22 rows for models we track and 34 other entries at the last read.
- Pass Last read November 18th, 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.