Which model leads on APEX-Agents?

Claude Sonnet 5.5 from Anthropic leads APEX-Agents with a score of 75.5. The median score across 56 tested models is 33.0.

Is APEX-Agents saturated?

No · the top score is 75.5 out of 100 (76%). There is still meaningful room for improvement on APEX-Agents.

Does APEX-Agents predict performance on other benchmarks?

Yes · APEX-Agents scores correlate 0.93 with ARC-AGI-2 across 38 shared models. Models that do well on APEX-Agents tend to do well on ARC-AGI-2.

How often is APEX-Agents data refreshed?

BenchGecko pulls updates daily. New model scores on APEX-Agents appear as soon as they are published by Epoch AI or the model provider.

Benchmark · AgentCompetitive

APEX-Agents

Name: APEX-Agents Benchmark
Creator: BenchGecko
License: https://creativecommons.org/licenses/by/4.0/

APEX-Agents · evaluates AI agents on complex, multi-step tasks requiring planning, tool use, and autonomous decision-making in realistic environments.

Updated 2026-09-28

Models tested

Top score

75.5

Claude Sonnet 5.5

Median

33.0

min 1.1

Top-5 spread

σ 3.7

Competitive

The Frontier

Best score over time · one chart, every benchmark

Chart type

Frontier on APEX-Agents rose from 1.1 to 75.5 in 23 months · +74.4 points · latest leader Claude Sonnet 5.5 from Anthropic.

Pink dots = frontier records · 13 totalClick to open model page

Full rankings

56 models tested · sorted by score

#	Model	Score	Price
1	Claude Sonnet 5.5· Anthropic	75.5	$2.00
2	Claude Opus 5.5· Anthropic	73.5	$4.00
3	Claude Fable 5.1· Anthropic	68.6	$10.00
4	Gemini 3.7 Flash· Google DeepMind	67.8	$0.75
5	Claude Opus 5· Anthropic	65.8	$5.00
6	Grok 4.6· xAI	65.3	$2.00
7	GPT-6 Astra· OpenAI	64.7	$10.00
8	Gemini 3.8 Flash· Google DeepMind	64.3	$0.75
9	Claude Fable 5· Anthropic	63.6	$10.00
10	GPT-5.6 Terra· OpenAI	58.2	$2.00
11	GLM 5.3· z-ai	56.6	$1.40
12	Grok 4.5· xAI	56.2	$2.00
13	Claude Sonnet 5· Anthropic	54.5	$2.00
14	GLM 5.3 Flash· z-ai	52.8	$0.15
15	GPT-5.4· OpenAI	52.4	$2.50
16	GPT-5.6 Sol· OpenAI	51.4	$2.00
17	Kimi K3· moonshotai	50.6	$0.67
18	Claude Opus 4.7· Anthropic	49.2	$5.00
19	Claude Opus 4.8· Anthropic	48.9	$5.00
20	DeepSeek V4 Pro 0813· DeepSeek	47.3	$0.55
21	Gemini 3.6 Flash· Google DeepMind	46.9	$0.75
22	Claude Opus 4.6· Anthropic	46.3	$5.00
23	Claude Sonnet 4.6· Anthropic	43.0	$3.00
24	GLM 5.1· z-ai	40.9	$0.97
25	MiniMax M3· minimax	37.7	$0.30
26	Kimi K2.7 Code· moonshotai	37.6	$0.61
27	Gemini 3.1 Pro Preview· Google DeepMind	35.3	$2.00
28	GPT-5.2· OpenAI	34.3	$1.75
29	GPT-5.3-Codex· OpenAI	31.7	$1.75
30	Gemini 3.5 Flash· Google DeepMind	27.5	$1.50
31	Qwen3.5 397B A17B· Alibaba Qwen	24.9	$0.55
32	GPT-5.4 Mini· OpenAI	24.6	$0.75
33	Gemini 3 Flash Preview· Google DeepMind	24.0	$0.50
34	Claude Opus 4.5· Anthropic	18.4	$5.00
35	Gemini 3 Pro· Google DeepMind	18.4	n/a
36	GPT-5· OpenAI	18.3	$1.25
37	GPT-5.1· OpenAI	17.5	$1.25
38	GLM 5· z-ai	17.2	$0.60
39	o3· OpenAI	17.2	$2.00
40	GPT-5.4 Nano· OpenAI	16.9	$0.20
41	Grok 4· xAI	15.2	n/a
42	Kimi K2.5· moonshotai	14.4	$0.45
43	Qwen 3.5 Plus (hosted 397B-A17B)· Alibaba Qwen	13.6	n/a
44	Claude Sonnet 4· Anthropic	9.3	$3.00
45	Claude Haiku 4.5· Anthropic	8.9	$1.00
46	DeepSeek V3.2· DeepSeek	7.0	$0.28
47	Gemini 2.5 Pro· Google DeepMind	6.6	$1.25
48	MiniMax M2.5· minimax	6.2	$0.27
49	gpt-oss-120b· OpenAI	4.4	$0.04
50	Kimi K2 Thinking· moonshotai	4.0	$0.60
51	GLM 4.7· z-ai	3.1	$0.60
52	GLM 4.6· z-ai	3.0	$0.43
53	Grok 3· xAI	2.1	n/a
54	Gemini 2.5 Flash· Google DeepMind	1.8	$0.30
55	GPT-4o (2024-11-20)· OpenAI	1.1	$2.50
56	o1· OpenAI	1.1	$15.00