Task ranking
Best LLM for Vision & Documents
Gemini 3.8 Flash is the provisional first pick for Vision & Documents as of November 15, 2026, while independent task measurements are still limited.[1][2][3][4][5] Ranks #2 on LMArena Vision's OCR and diagram categories, and achieved S-tier (0-2 errors across 11 runs) on a tax form reading benchmark. Claude Opus 5.5 scores within two points, so read the top two as close rather than as a measured gap. Watch out: Watch for chunking artifacts in mixed OCR mode: a 5-page PDF produced 16 page headers (10 marked 'unknown') and 92 product entries with 18 duplicates containing contradictory field values. Claude Opus 5.5 is the next-ranked alternative. It ranks #3 of 74 on LMArena Vision's diagram category (Elo 1339), based on blind human preference for interpreting charts and diagrams.
About this recommendation
- Updated
- Evidence through
- Sources
- 13
- Revision
- v132
Decision audit
Why this result
Inspect the inputs and the computed order behind the recommendation.
Models screened
24
live candidates
Evaluation feeds
6
task-weighted
Winner coverage
40%
intended feed weight
Largest provider share
3 of 8
Google: Gemini 3.8 Flash leads the measured task comparison with limited independent task measurements. Community volume, release age and usage popularity add no ranking points. The leading scores are close under this policy.
Sources evaluated
These are the weights used in scoring. Missing measurements are omitted from each model’s average. A feed with no result for any model released in the last 60 days is stale and does not score until it catches up.
| Evaluation feed | Weight | Winner result | Field measured |
|---|---|---|---|
| VLMEvalKit tasksdisabled; last successful check unknownunavailable | 0% | feed unavailable | 0/24 |
| LMArena DocumentObserved 2026-11-09 · evaluation date not supplied · high effortstale | 0% | no current result | 0/24 |
| LMArena VisionObserved 2026-11-09 · evaluation date not supplied · high effort | 37% | #7 | 24/24 |
| LMArena Vision OCRObserved 2026-11-09 · evaluation date not supplied · high effort | 37% | #2 | 24/24 |
| Structured-output evaldisabled; last successful check unknownunavailable | 0% | feed unavailable | 0/24 |
| LMArena Vision DiagramsObserved 2026-11-09 · evaluation date not supplied · high effort | 25% | #2 | 24/24 |
Provider concentration
Each exact model is scored separately; provider identity is not a ranking input.
- Google3 models
- Meta2 models
- OpenAI2 models
- Anthropic1 model
Decision table
Every published model is shown in computed order. Practitioner sources are distinct community threads, not the citations repeated in the prose below.
Google · coverage 40% · provisional evidence
#2 LMArena Vision Diagrams · #2 LMArena Vision OCR
Anthropic · coverage 40% · close score · provisional evidence
#3 LMArena Vision Diagrams · #4 LMArena Vision
Meta · coverage 40% · provisional evidence
#5 LMArena Vision Diagrams · #6 LMArena Vision OCR
Google · coverage 40% · provisional evidence
#5 LMArena Vision · #7 LMArena Vision OCR
OpenAI · coverage 40% · provisional evidence
#6 LMArena Vision Diagrams · #7 LMArena Vision
Meta · coverage 40% · provisional evidence
#5 LMArena Vision · #8 LMArena Vision OCR
OpenAI · coverage 40% · provisional evidence
#14 LMArena Vision · #16 LMArena Vision OCR
Google · coverage 40% · provisional evidence
#13 LMArena Vision OCR · #16 LMArena Vision Diagrams
Task score averages the available, task-weighted measurements. Benchmark positions compare models matched to our catalog. Models measured by at least two independent evaluator families are considered first. Missing tests, practitioner report counts, popularity and release age add no points. Coverage describes evidence breadth, not a probability of correctness. Scores within two points are shown as close under this policy; this is not a statistical confidence interval. When the leading scores are that close, the model measured by more evaluator families ranks first.
- 01
Gemini 3.8 FlashGoogle · $0.83 in / $4.13 out per M tokens · 1.0M context
Ranks #2 on LMArena Vision's OCR and diagram categories, and achieved S-tier (0-2 errors across 11 runs) on a tax form reading benchmark.
Best when: Use for OCR-heavy workflows: ranks #2 of 74 on LMArena Vision's OCR category with Elo 1315, and #2 of 74 on diagrams with Elo 1342.More tips
- Deploy for document extraction with handwritten content: successfully extracted unstructured shipyard PDFs including handwritten notes and non-standard tables.
- Prioritize for structured document parsing: scored 0 to 2 wrong across 11 runs on tax form reading, placing it in S-tier above Qwen3-VL and olmOCR pipelines.
Watch out for
- Watch for chunking artifacts in mixed OCR mode: a 5-page PDF produced 16 page headers (10 marked 'unknown') and 92 product entries with 18 duplicates containing contradictory field values.
- 02
Claude Opus 5.5Anthropic · $4 in / $20 out per M tokens · 1M context
Claude Opus 5.5 ranks #3 of 74 on LMArena Vision's diagram category (Elo 1339), based on blind human preference for interpreting charts and diagrams.
Best when: Its strongest result for this task is on LMArena Vision diagrams, which measures interpreting charts and diagrams. - 03
Muse Spark 1.3Meta · $1.25 in / $4.25 out per M tokens · 1.0M context
Muse Spark 1.3 ranks #5 of 74 on LMArena Vision's diagram category (Elo 1329), based on blind human preference for interpreting charts and diagrams.
Best when: Its strongest result for this task is on LMArena Vision diagrams, which measures interpreting charts and diagrams. - 04
Gemini 3.7 FlashGoogle · $0.75 in / $3.75 out per M tokens · 1.0M context
Ranks #5 on LMArena's vision arena with balanced performance across document, OCR, and diagram tasks.
Best when: Use as a general-purpose vision model: ranks #5 of 79 on LMArena's vision arena with Elo 1294.More tips
- Deploy for document workflows: ranks #10 of 36 on LMArena's document arena with score 1476.
No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 05
GPT-6.1 SolOpenAI · $2 in / $10 out per M tokens · 1.1M context
Tied at #7 on LMArena's vision arena with strong diagram performance but weaker OCR ranking.
Best when: Use for chart and diagram analysis: ranks #6 of 74 on LMArena Vision's diagram category with Elo 1327.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 06
Muse Spark 1.2Meta · $1.25 in / $4.25 out per M tokens · 1.0M context
Muse Spark 1.2 ranks #5 of 79 on LMArena's vision arena (Elo 1294), based on human preference on image-understanding tasks.
Best when: Its strongest result for this task is on LMArena Vision, which measures reading images, screenshots, and documents.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 07
GPT-6 AstraOpenAI · $10 in / $50 out per M tokens · 1.1M context
GPT-6 Astra ranks #14 of 79 on LMArena's vision arena (Elo 1283), based on human preference on image-understanding tasks.
Best when: Its strongest result for this task is on LMArena Vision, which measures reading images, screenshots, and documents.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
- 08
Gemini 3.6 FlashGoogle · $0.75 in / $3.75 out per M tokens · 1.0M context
Gemini 3.6 Flash ranks #13 of 74 on LMArena Vision's OCR category (Elo 1298), based on blind human preference for reading text out of images.
Best when: Its strongest result for this task is on LMArena Vision OCR, which measures reading text out of images and scans.No practitioner reports are linked yet, so no cautions are listed. Absence of reports is not evidence of reliability.
Frequently asked
- What is the best model for Vision & Documents?
- Gemini 3.8 Flash ranks first as of November 15, 2026. Ranks #2 on LMArena Vision's OCR and diagram categories, and achieved S-tier (0-2 errors across 11 runs) on a tax form reading benchmark.[1][2][3]
- What should I watch out for with Gemini 3.8 Flash?
- Watch for chunking artifacts in mixed OCR mode: a 5-page PDF produced 16 page headers (10 marked 'unknown') and 92 product entries with 18 duplicates containing contradictory field values.[4]
- What is the best alternative to Gemini 3.8 Flash?
- Claude Opus 5.5 ranks second. Claude Opus 5.5 ranks #3 of 74 on LMArena Vision's diagram category (Elo 1339), based on blind human preference for interpreting charts and diagrams.[5]
Sources
- 1
“Within the 74 catalog models matched to this evaluation: Ranks #2 of 74 on LMArena Vision's OCR category (Elo 1315), based on blind human preference for reading text out of images.”
LMArena Vision OCR · Benchmark · Nov 9, 2026 - 2
“Within the 74 catalog models matched to this evaluation: Ranks #2 of 74 on LMArena Vision's diagram category (Elo 1342), based on blind human preference for interpreting charts and diagrams.”
LMArena Vision diagrams · Benchmark · Nov 9, 2026 - 3
“Verdict (revised after the 2026-10-22 review). The board stands on our numbers, said as "on our test, October 2026": S Gemini 3.8 Flash (0 to 2 wrong across 11 runs), A Qwen3-VL 235B, B Qwen 3.8 27B, C olmOCR 2 + Gemini, D GLM 5.3 Flash, F granite-docling and Gemma 4 31B. Gemma's F is for the model as we ran it, and outside tests on simpler documents put it near the top, though still behind Gemini wherever both are tested. The closer still works said that way: a model that gets 97.6% of inv…”
herzo175 · GitHub · Oct 22, 2026 - 4
fix(knowledge): 混合 OCR 產生重複頁面與矛盾商品條目(5 頁 PDF 出現 16 組頁面標題、92/74 商品重複)
Original-language excerpt:“現象 評測租戶知識庫「DM」的 DM_家樂福_采集漫旅_商品頁_p11-15.pdf(5 頁)以 Gemini 3.8 Flash 走混合 OCR(切片 2x3 + 整頁)處理後,落庫內容出現: - 16 組 【頁面標題】(其中 10 組為「不詳」),但原始 PDF 只有 5 頁 - 92 筆 商品: 條目,只有 74 筆唯一 — 18 筆重複 - 重複條目的欄位互相矛盾,例如同一商品「鮮剖100%椰子汁」出現兩次: - 一處 品牌:半天水、規格:含椰肉/純天然 520毫升×3、原價:120~129 - 另一處 品牌:鮮剖、規格:520毫升×3、原價:120-129元、備註:含椰肉/純天然 - 前者所在區段有完整頁面標題「單筆購買第12~15頁…」,後者所在區段頁面標題為「不詳」 已排除 chunk overlap 造成的假重複:把 chunk 依 chunk_index 串接後長度等於各 chunk 長度總和(無重疊),上述統計是文件本體的真實內容。 也已排除 worker 跑舊版:VM `poc-rag-v…”
larry610881 · GitHub · Sep 8, 2026 - 5
“Within the 74 catalog models matched to this evaluation: Ranks #3 of 74 on LMArena Vision's diagram category (Elo 1339), based on blind human preference for interpreting charts and diagrams.”
LMArena Vision diagrams · Benchmark · Nov 9, 2026 - 6
“Overview Upgrade scripts/prototype_experiment.py from local regex parsing to a live multimodal pipeline utilizing the Gemini 3.8 Flash API, automating the extraction of unstructured shipyard PDFs (including handwritten notes and non-standard tables) and drafting Lloyd's-standard Preliminary Survey Reports from deterministic appraisal outputs. Context & Motivation - Shipyard repair specifications vary widely in layout, trade nomenclature, and often include handwritten field revisions…”
yohei1126 · GitHub · Sep 26, 2026 - 7
“Within the 74 catalog models matched to this evaluation: Ranks #5 of 74 on LMArena Vision's diagram category (Elo 1329), based on blind human preference for interpreting charts and diagrams.”
LMArena Vision diagrams · Benchmark · Nov 9, 2026 - 8
“Within the 79 catalog models matched to this evaluation: Ranks #5 of 79 on LMArena's vision arena (Elo 1294), based on human preference on image-understanding tasks.”
LMArena vision arena · Benchmark · Nov 9, 2026 - 9
“Within the 36 catalog models matched to this evaluation: Ranks #10 of 36 on LMArena's document arena (score 1476), based on blind preference for document tasks.”
LMArena document arena · Benchmark · Nov 9, 2026 - 10
“Within the 74 catalog models matched to this evaluation: Ranks #6 of 74 on LMArena Vision's diagram category (Elo 1327), based on blind human preference for interpreting charts and diagrams.”
LMArena Vision diagrams · Benchmark · Nov 9, 2026 - 11
“Within the 79 catalog models matched to this evaluation: Ranks #5 of 79 on LMArena's vision arena (Elo 1294), based on human preference on image-understanding tasks.”
LMArena vision arena · Benchmark · Nov 9, 2026 - 12
“Within the 79 catalog models matched to this evaluation: Ranks #14 of 79 on LMArena's vision arena (Elo 1283), based on human preference on image-understanding tasks.”
LMArena vision arena · Benchmark · Nov 9, 2026 - 13
“Within the 74 catalog models matched to this evaluation: Ranks #13 of 74 on LMArena Vision's OCR category (Elo 1298), based on blind human preference for reading text out of images.”
LMArena Vision OCR · Benchmark · Nov 9, 2026
Rankings synthesized from community evidence and open benchmarks. See methodology. Not driven by vendor marketing.