Which model leads ChartQA Test?

Q: Which model leads ChartQA Test?

InternVL3_5-4B tops the board at 86.2

ChartQA Test Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensofficial card harness

The field · 13 measured

42.186.2

13 models measured, most on the official card harness. InternVL3_5-4B tops the board at 86.2.

13 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 13Bars show 42.1–86.2 · ◆ solid = first-party or better · outlined = arm's-length

01InternVL3_5-4BOpenGVLab86.2———

02Qwen3.5 4BAlibaba84.2$0.03/M$0.15/M$0.001

03Phi 3.5 Vision Instruct+1 altMicrosoft82.1———

04InternVL3_5-2BOpenGVLab81.7———

05LFM2.5-VL-3BLiquid AI81.3———

06North-Micro-Vision-Instruct+1 altCohere80.8———

07LFM2-VL-3BLiquid AI80.4———

08Ministral 3 3B+1 altMistral79.1$0.10/M$0.10/M$0.001

09Qwen3.5 2B+2 altsAlibaba78.4———

10LFM2.5-VL-1.6B+1 altLiquid AI73.9———

11Qwen3 VL 2B Instruct+1 altAlibaba69.3———

12Gemma 4 E2B+2 altsGoogle43.2———

13Gemma 4 E4BGoogle42.1$0.02/M$0.10/M$0.001

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 13 models scored · 0 independently verified · 13 vendor-reported · 0 with source disagreement. How these tiers are assigned