Vision Rankings

This page shows Arena AI A daily snapshot of the top 30 on the leaderboard, making it easy to quickly understand recent model performance. For the full leaderboard, filters, and latest changes, please visit the Arena AI official website.

This leaderboard evaluates the performance of multimodal models on image understanding tasks.

Data updated at: 2026-09-01 13:25:31 UTC / 2026-09-01 21:25:31 CST (Beijing time)

The leaderboard reflects specific evaluations and user voting preferences, and does not mean a model will necessarily perform better on your tasks. When choosing a model, you should also consider price, speed, context, tool calling, privacy, and regional availability.

Top 30

Rank	Rank range	Model	Score	Votes	Price $/million tokens	Context
1	1-6	claude-fable-5 1	1313 (±8)	10,002	$10 / $50	1M
2	1-11	claude-opus-4-7-high 1	1301 (±7)	21,137	$5 / $25	1M
3	1-15	qwen3.8-max 1	1300 (±8)	7,244	$2 / $6	1M
4	1-14	claude-opus-4-7 1	1299 (±7)	21,457	$5 / $25	1M
5	1-14	claude-opus-4-6-high 1	1299 (±7)	20,883	$5 / $25	1M
6	2-25	muse-spark 1	1294 (±9)	5,585	N/A	N/A
7	2-24	claude-opus-4-6 1	1293 (±7)	25,210	$5 / $25	1M
8	1-29	muse-spark-1.2 (xHigh) 1	1292 (±15)	1,844	$1.25 / $4.25	N/A
9	2-26	claude-opus-5-high 1	1290 (±8)	8,271	$5 / $25	1M
10	2-27	gemini-3-pro 1	1289 (±8)	13,019	$2 / $12	1M
11	3-27	gpt-5.5 1	1286 (±7)	21,389	$5 / $30	1.1M
12	3-28	claude-opus-4-8-high 1	1285 (±7)	13,521	$5 / $25	1M
13	2-31	gemini-3.6-flash-high 1	1285 (±11)	3,816	$0.75 / $3.75	1M
14	3-29	gemini-3.5-flash-high 1	1284 (±8)	8,259	$0.75 / $4.50	1M
15	6-29	gpt-5.5-high 1	1284 (±7)	20,121	$5 / $30	1.1M
16	6-31	gemini-3.5-flash-medium 1	1283 (±8)	8,261	$0.75 / $4.50	1M
17	5-32	gpt-5.6-sol-xhigh 1	1282 (±9)	5,837	$4 / $20	N/A
18	6-32	grok-4.5 1	1282 (±9)	6,401	$2 / $6	500K
19	6-31	gpt-5.4-high 1	1281 (±7)	23,263	$1.25 / $7.50	1.1M
20	6-32	gpt-5.4 1	1281 (±7)	21,245	$1.25 / $7.50	1.1M
21	6-32	claude-opus-4-8 1	1279 (±7)	13,980	$5 / $25	1M
22	6-32	muse-spark-1.1 1	1279 (±9)	6,769	$1.25 / $4.25	1M
23	7-32	gpt-5.2-chat-latest-20260210 1	1279 (±7)	15,446	$1.75 / $14	128K
24	8-32	gemini-3.1-pro-preview 1	1278 (±6)	39,412	$1 / $6	1M
25	6-33	gpt-5.5-instant 1	1278 (±9)	7,225	$5 / $30	1.1M
26	9-33	claude-sonnet-4-6 1	1275 (±6)	25,610	$1.50 / $7.50	1M
27	6-42	glm-5.3-flash 1	1273 (±17)	1,389	$0.15 / $0.50	1M
28	11-34	gemini-3-flash 1	1272 (±5)	36,747	$0.50 / $3	1M
29	15-39	claude-sonnet-5-high 1	1267 (±8)	8,522	$2 / $10	1M
30	12-42	gemini-3.5-flash-lite 1	1266 (±11)	3,886	$0.15 / $1.25	1M

How to read this table

• Rank / rank range: The relative rank estimated by Arena AI based on head-to-head votes; the rank range indicates the variation in rank within the confidence interval.
• Score: Relative score, suitable for comparing models on the leaderboard at the same time.
• Votes / sessions: Sample size reference; with fewer samples, rankings are usually more likely to fluctuate.
• Price $/million tokens: Reference price per million input / output tokens.
• Context: The maximum context length supported by the model.

Three things to confirm when choosing a model

1. Whether the provider actually offers this model, and whether it is available in your region and account;
2. Whether the API price, rate limits, and context fit your task;
3. Run a small test with 3 to 5 real tasks; do not rely only on the overall leaderboard rank.

Data sources

Data comes from Arena AI Official Vision Leaderboard, updated daily by GitHub Actions. For model pricing, licenses, and capabilities, please refer to the official information from the model provider.