Browser Use Bench v2.1 | 180 tasks | October 17, 2026

Agent benchmarks.

Name: Browser Use Bench v2.1 model results
Creator: Browser Use

GPT-6.1 Sol medium scores highest at 79.6. Browser Use's own agent scores 72.9 at 11¢ a task, level with Claude Sonnet 5.5 high at 14× less.

Agents Browsers Consumer tasks

OpenAI
Anthropic
xAI
DeepSeek
Z.ai
Browser Use

#	Model	Score	95% range	Cost / task	Median time
1	GPT-6.1 Sol medium	79.6	76–83	$1.30	8 min
2	GPT-6 Astra medium	77.9	74–82	$10.77	9 min
3	DeepSeek V4.1 Flash 1017 max	75.0	71–79	22¢	17 min†
3	Claude Opus 5.5 high	75.0	71–79	$2.35	12 min
5	GPT-6 Sol medium	74.4	71–78	$1.02	6 min
6	DeepSeek V4.1 Flash 1017 low	74.0	70–78	12¢	12 min†
7	DeepSeek V4.1 Flash 1017 high	73.6	70–77	15¢	13 min†
8	Grok 4.7 high	73.3	70–77	$3.46	18 min
9	Browser Use (BU fast)	72.9	69–77	11¢	9 min
10	Claude Sonnet 5.5 high	72.7	69–77	$1.53	9 min
11	Claude Fable 5.1 high	72.0	68–76	$5.19	17 min
12	DeepSeek V4.1 Flash max	70.9	67–75	18¢	18 min
13	GPT-5.6 Luna xhigh	69.6	66–73	28¢	15 min
14	Claude Haiku 5.5 xhigh	69.5	65–73	46¢	14 min
15	GPT-6 Luna xhigh	69.0	65–73	16¢	11 min
16	Claude Opus 5	66.8	62–71	$4.72	19 min
17	Claude Haiku 5.5 max	65.3	61–70	$1.07	21 min
18	GPT-6 Luna medium	64.3	60–68	7.7¢	8 min
18	Claude Haiku 5.5 high	64.3	60–68	23¢	9 min
20	GLM 5.3 FlashX max	59.7	55–64	55¢	14 min
21	Claude Haiku 5.5 medium	53.4	50–57	9.2¢	5 min
22	Claude Haiku 5.5 low	45.6	42–50	2.7¢	2 min

† Ran on different CI runners from the rest of the lineup, so its time is not comparable.

180 long tasks on live websites, the same for every model, run in the open-source BrowserCode harness (Browser Use's own agent runs in ours). Mean weighted rubric score per task, 0-100. Estimated model cost per task in USD. Excludes search, judge, browser and runner costs. Runner and encrypted tasks: github.com/browser-use/benchmark.

Each model

Run your tasks in the cloud

›Earlier benchmarks: Internal Bench Hard, OnlineMind2Web, BU Bench V1

Internal Bench Hard

Browser Use: 82% of tasks solved at 17¢ each. That is 20 points better than Opus 5, which costs 20× more per solved task.

Toolstrict accuracyCost

Browser Use82%$0.17

Opus 562%$3.40

Gemini 3.1 Pro59%$2.20

Sonnet 559%$1.55

GPT-5.652%$1.10

Gemini 3.6 Flash46%$0.62

GPT-537%$0.44