IFBench

Artificial Analysis · 50 models

Rank	Model	Score
1	Grok 4.3 (Medium)	0.8333
2	Grok 4.20 0309 (Reasoning)	0.8293
3	MiniMax-M3	0.8286
4	Nemotron 3 Ultra 550B A55B (Reasoning)	0.8136
5	Grok 4.3 (High)	0.8129
6	Grok 4.20 0309 v2 (Reasoning)	0.8122
7	Grok 4.3 (Low)	0.8095
8	Qwen3.7 Max	0.8054
9	Nemotron Cascade 2 30B A3B	0.8041
10	MiMo-V2.5-Pro (Reasoning)	0.7986
11	Nova 2.0 Pro Preview (Low)	0.7959
12	DeepSeek V4 Flash 0420 (Max)	0.7918
13	Nova 2.0 Pro Preview (Medium)	0.7905
14	Qwen3.5 397B A17B (Reasoning)	0.7878
15	Gemini 3 Flash Preview (Reasoning)	0.7796
15	Qwen3.7 Plus	0.7796
17	GPT-5.2 Codex (Xhigh)	0.7762
18	Gemini 3.1 Flash-Lite	0.7721
19	Gemini 3.1 Pro Preview	0.7714
20	Qwen3.6 Max Preview	0.766
21	DeepSeek V4 Pro 0424 (Max)	0.7646
22	Gemini 3.5 Flash (High)	0.7633
23	GLM-5.1 (Reasoning)	0.7626
24	Kimi K2.6 (Reasoning)	0.7599
25	GPT-5.4 nano (Xhigh)	0.7592
25	Muse Spark	0.7592
27	GPT-5.5 (Xhigh)	0.7585
28	MiniMax-M2.7	0.7571
28	Qwen3.5 122B A10B (Reasoning)	0.7571
30	Gemma 4 31B (Reasoning)	0.7558
30	Qwen3.5 27B (Reasoning)	0.7558
32	GPT-5 mini (High)	0.7544
32	GPT-5.2 (Xhigh)	0.7544
34	GPT-5.3 Codex (Xhigh)	0.7537
35	Qwen3.6 Plus	0.7517
36	Gemini 3.5 Flash (Medium)	0.7456
37	GPT-5 Codex (High)	0.7415
38	Command A+	0.7395
38	GPT-5.4 (Xhigh)	0.7395
40	Gemma 4 12B (Reasoning)	0.7354
41	DeepSeek V4 Flash 0420 (High)	0.7347
42	GLM-5.2 (Max)	0.7333
43	GPT-5.4 mini (Xhigh)	0.7327
44	GLM-5-Turbo	0.732
45	GPT-5 (High)	0.7306
46	GPT-5.1 (High)	0.7286
47	GPT-5.6 Sol (Max)	0.7265
48	Qwen3.5 35B A3B (Reasoning)	0.7252
49	Gemma 4 26B A4B (Reasoning)	0.7245
50	GLM-5 (Reasoning)	0.7231