AI Model Rankings

Measured success rate and latency for the models we serve, from real API calls made through ApexApi during health sweeps over the last 30 days. These are our own measured numbers, not vendor claims.

#	Model	Checks	Success	Avg latency	Last checked
1	Gemini 3.1 Flash Lite google	1	100%	1.8s	Aug 30
2	Gemini 3.1 Flash Lite Preview google	1	100%	1.9s	Aug 30
3	GPT-4.1 Mini openai	1	100%	2.4s	Aug 30
4	Gemini 3.6 Flash google	1	100%	2.6s	Aug 30
5	GPT-4.1 openai	1	100%	2.7s	Aug 30
6	Ministral 3 8B 2512 mistral	1	100%	2.8s	Aug 30
7	Ministral 3 3B 2512 mistral	1	100%	2.8s	Aug 30
8	Mistral Medium 3 mistral	1	100%	2.8s	Aug 30
9	Gemini 2.5 Flash Lite google	1	100%	2.9s	Aug 30
10	GPT-OSS 20B bedrock	1	100%	2.9s	Aug 30
11	Claude Haiku 4.5 anthropic	1	100%	2.9s	Aug 30
12	GPT-5.6 Luna openai	1	100%	3.0s	Aug 30
13	Gemini 3.5 Flash Lite google-vertex	1	100%	3.0s	Aug 30
14	GPT-5.5 openai	1	100%	3.1s	Aug 30
15	GPT-5.4 Mini openai	1	100%	3.2s	Aug 30
16	GPT-4o openai	1	100%	3.2s	Aug 30
17	GPT-5.1 openai	1	100%	3.2s	Aug 30
18	Mistral Small 4 mistral	1	100%	3.3s	Aug 30
19	Gemini 3 Flash Preview google	1	100%	3.3s	Aug 30
20	Gemini 3.5 Flash google	1	100%	3.3s	Aug 30
21	GLM 4.7 Flash bedrock	1	100%	3.3s	Aug 30
22	o4 Mini openai	1	100%	3.3s	Aug 30
23	GPT-4o openai	1	100%	3.3s	Aug 30
24	GPT-4o openai	1	100%	3.3s	Aug 30
25	o3 openai	1	100%	3.4s	Aug 30
26	Claude Haiku 4.5 anthropic-vertex	1	100%	3.4s	Aug 30
27	GPT-5.6 Terra openai	1	100%	3.4s	Aug 30
28	GPT-4o-mini openai	1	100%	3.5s	Aug 30
29	Claude Opus 4.5 anthropic-vertex	1	100%	3.5s	Aug 30
30	GPT-OSS 120B bedrock	1	100%	3.5s	Aug 30
31	Claude Opus 4.8 anthropic-vertex	1	100%	3.5s	Aug 30
32	Kimi K2.5 bedrock	1	100%	3.5s	Aug 30
33	GPT-4o openai	1	100%	3.5s	Aug 30
34	Nano Banana google	1	100%	3.5s	Aug 30
35	GPT-5.4 Nano openai	1	100%	3.5s	Aug 30
36	Ministral 3 8B 2512 bedrock	1	100%	3.6s	Aug 30
37	Mistral Medium 3.5 mistral	1	100%	3.6s	Aug 30
38	GPT-4o-mini openai	1	100%	3.6s	Aug 30
39	Devstral 2 2512 bedrock	1	100%	3.6s	Aug 30
40	GPT-5.4 openai	1	100%	3.6s	Aug 30
41	GPT-5 openai	1	100%	3.6s	Aug 30
42	Gemma 4 26B A4B google	1	100%	3.6s	Aug 30
43	Codestral 2508 mistral	1	100%	3.7s	Aug 30
44	Claude Opus 4.8 anthropic	1	100%	3.7s	Aug 30
45	GPT-5.2 openai	1	100%	3.7s	Aug 30
46	GPT-4.1 Nano openai	1	100%	3.7s	Aug 30
47	Claude Sonnet 4.6 anthropic	1	100%	3.7s	Aug 30
48	Ministral 3 14B 2512 mistral	1	100%	3.8s	Aug 30
49	GPT-3.5 Turbo openai	1	100%	3.8s	Aug 30
50	o3 Mini openai	1	100%	3.8s	Aug 30
51	Ministral 3 3B 2512 bedrock	1	100%	3.8s	Aug 30
52	GLM 4.7 bedrock	1	100%	3.9s	Aug 30
53	Qwen3 VL Flash alibaba	1	100%	4.0s	Aug 30
54	Ministral 3 14B 2512 bedrock	1	100%	4.0s	Aug 30
55	GPT-4 openai	1	100%	4.1s	Aug 30
56	Qwen3 VL Plus alibaba	1	100%	4.1s	Aug 30
57	Qwen3.6 Flash alibaba	1	100%	4.2s	Aug 30
58	o1 openai	1	100%	4.2s	Aug 30
59	Claude Opus 4.7 anthropic	1	100%	4.2s	Aug 30
60	Claude Opus 4.7 anthropic-vertex	1	100%	4.2s	Aug 30
61	Claude Opus 5 anthropic-vertex	1	100%	4.4s	Aug 30
62	Gemma 4 31B google	1	100%	4.5s	Aug 30
63	Claude Opus 4.6 anthropic	1	100%	4.5s	Aug 30
64	DeepSeek V4 Flash deepseek	1	100%	4.5s	Aug 30
65	GPT-5 Mini openai	1	100%	4.6s	Aug 30
66	Claude Opus 4.5 anthropic	1	100%	4.6s	Aug 30
67	Gemini 3.1 Pro Preview Custom Tools google	1	100%	4.7s	Aug 30
68	GPT-5.6 Sol openai	1	100%	4.7s	Aug 30
69	Qwen3 Coder Plus alibaba	1	100%	4.7s	Aug 30
70	GPT-4 Turbo openai	1	100%	4.9s	Aug 30
71	DeepSeek V4 Pro deepseek	1	100%	4.9s	Aug 30
72	Gemini 3.1 Pro Preview google	1	100%	4.9s	Aug 30
73	Claude Sonnet 4.6 anthropic-vertex	1	100%	5.0s	Aug 30
74	GPT-5 Nano openai	1	100%	5.1s	Aug 30
75	Qwen3.5 Flash alibaba	1	100%	5.3s	Aug 30
76	Qwen3 Max alibaba	1	100%	5.5s	Aug 30
77	GPT-3.5 Turbo 16k openai	1	100%	5.5s	Aug 30
78	Claude Opus 4.6 anthropic-vertex	1	100%	5.6s	Aug 30
79	Gemini 2.5 Flash google	1	100%	5.8s	Aug 30
80	Qwen3.7 Max alibaba	1	100%	6.1s	Aug 30
81	GLM 5 bedrock	1	100%	6.8s	Aug 30
82	Qwen3.7 Plus alibaba	1	100%	8.0s	Aug 30
83	Gemini 2.5 Pro google	1	100%	9.4s	Aug 30
84	Kimi K2 Thinking bedrock	1	100%	9.4s	Aug 30
85	MiniMax M2.5 bedrock	1	100%	12.6s	Aug 30
86	Devstral 2 2512 mistral	1	100%	15.2s	Aug 30
87	Qwen3.5 Plus alibaba	1	100%	87.0s	Aug 30
88	GPT-4o Search Preview openai	1	0%	—	Aug 30
89	Mistral Large 3 2512 bedrock	1	0%	—	Aug 30
90	GPT-4o-mini Search Preview openai	1	0%	—	Aug 30
91	Mistral Large 3 2512 mistral	1	0%	—	Aug 30

Latency is the average of successful calls and depends on prompt size, region, and upstream load. Success rate reflects our smoke tests, not a guarantee. Call any of these through one API and key at the model catalog.