Skip to main content
Meo

IFBench — AI model leaderboard

AI models ranked by IFBench, an aggregated third-party benchmark from artificial_analysis. Higher is better. Cross-referenced against our first-party meo scores and Effective Value (𝕍).

Ranking 79 models across the full field · as of 2026-06-07.

#ModelLabIFBench
1MiniMax: MiniMax M3minimax82.9%
2xAI: Grok 4.3x-ai81.3%
3Qwen: Qwen3.7 Maxqwen80.5%
4Xiaomi: MiMo-V2.5-Proxiaomi79.9%
5DeepSeek: DeepSeek V4 Flashdeepseek79.2%
6Qwen: Qwen3.5 397B A17Bqwen78.8%
7Qwen: Qwen3.7 Plusqwen78.0%
8OpenAI: GPT-5.2-Codexopenai77.6%
9Google: Gemini 3.1 Flash Litegoogle77.2%
10Google: Gemini 3.1 Pro Previewgoogle77.1%
11DeepSeek: DeepSeek V4 Prodeepseek76.5%
12Google: Gemini 3.5 Flashgoogle76.3%
13Z.ai: GLM 5.1z-ai76.3%
14OpenAI: GPT-5.4 Nanoopenai75.9%
15OpenAI: GPT-5.5openai75.9%
16MiniMax: MiniMax M2.7minimax75.7%
17Qwen: Qwen3.5-122B-A10Bqwen75.7%
18Google: Gemma 4 31Bgoogle75.6%
19OpenAI: GPT-5.2 Chatopenai75.4%
20OpenAI: GPT-5 Miniopenai75.4%
21OpenAI: GPT-5.3-Codexopenai75.4%
22Qwen: Qwen3.6 Plusqwen75.2%
23OpenAI: GPT-5 Codexopenai74.1%
24OpenAI: GPT-5.4openai73.9%
25OpenAI: GPT-5.4 Miniopenai73.3%
26Z.ai: GLM 5 Turboz-ai73.2%
27OpenAI: GPT-5openai73.1%
28OpenAI: GPT-5.1openai72.9%
29Google: Gemma 4 26B A4B (free)google72.4%
30OpenAI: o3openai71.4%
31Upstage: Solar Pro 3upstage71.2%
32OpenAI: o1openai70.3%
33OpenAI: GPT-5.1-Codexopenai70.0%
34Inception: Mercury 2inception69.8%
35OpenAI: gpt-oss-120bopenai69.0%
36Mistral: Mistral Medium 3.5mistralai68.8%
37OpenAI: o4 Miniopenai68.7%
38OpenAI: GPT-5.1-Codex-Miniopenai67.9%
39Qwen: Qwen3.6 27Bqwen67.6%
40OpenAI: GPT-5 Nanoopenai67.6%
41StepFun: Step 3.7 Flashstepfun67.3%
42OpenAI: o3 Mini Highopenai67.1%
43Qwen: Qwen3.5-9Bqwen66.7%
44Kwaipilot: KAT-Coder-Pro V2kwaipilot66.7%
45StepFun: Step 3.5 Flashstepfun66.5%
46OpenAI: gpt-oss-20bopenai65.1%
47Qwen: Qwen3.6 35B A3Bqwen64.4%
48Tencent: Hy3 previewtencent63.1%
49Anthropic: Claude Opus 4.8anthropic62.2%
50Z.ai: GLM 5V Turboz-ai61.1%
51Anthropic: Claude Opus 4.7anthropic58.6%
52inclusionAI: Ling-2.6-flashinclusionai57.4%
53inclusionAI: Ling-2.6-1Tinclusionai56.9%
54Arcee AI: Trinity Large Thinkingarcee-ai56.3%
55Google: Gemini 3 Flash Previewgoogle55.1%
56Google: Gemini 2.5 Progoogle48.7%
57inclusionAI: Ring-2.6-1Tinclusionai44.6%
58OpenAI: GPT-4.1openai43.0%
59Meta: Llama 4 Maverickmeta-llama43.0%
60Google: Gemini 2.5 Flash Lite Preview 09-2025google41.8%
61Anthropic: Claude Sonnet 4.6anthropic41.2%
62Xiaomi: MiMo-V2-Flashxiaomi39.9%
63Meta: Llama 4 Scoutmeta-llama39.5%
64Google: Gemini 2.5 Flashgoogle39.0%
65IBM: Granite 4.1 8Bibm-granite38.6%
66OpenAI: GPT-4.1 Miniopenai38.3%
67Google: Gemma 3 12Bgoogle36.7%
68Cohere: Command Acohere36.5%
69OpenAI: GPT-4o (2024-08-06)openai36.0%
70Qwen: Qwen3 Coder Nextqwen35.2%
71OpenAI: GPT-4oopenai34.3%
72Prime Intellect: INTELLECT-3prime-intellect34.0%
73OpenAI: GPT-4.1 Nanoopenai32.0%
74Google: Gemma 3 27Bgoogle31.8%
75OpenAI: GPT-4o-miniopenai31.0%
76Reka Flash 3rekaai30.4%
77Google: Gemma 3 4Bgoogle28.3%
78Microsoft: Phi 4microsoft23.5%
79Microsoft: Phi 4 Mini Instructmicrosoft21.1%

Artificial Analysis (artificialanalysis.ai). Redistribution requires an AA commercial license.

← All rankingsMethodology & 𝕍 →