LLM Leaderboard
Ranked by GPQA Diamond · source AA
Graduate-level science Q&A. Strong reasoning signal.
#1 by GPQA Diamond
GPT-5.6 Sol (max)
94.1%
OpenAI · in $5.00 / out $30.0 per 1M · 65 tok/s
| # | Model | GPQA |
|---|---|---|
| 1 | GPT-5.6 Sol (max) | 94.1% |
| 2 | Gemini 3.1 Pro Preview | 94.1% |
| 3 | Kimi K3 | 93.5% |
| 4 | GPT-5.5 (xhigh) | 93.5% |
| 5 | GPT-5.5 (high) | 93.2% |
| 6 | GPT-5.6 Sol (xhigh) | 93.1% |
| 7 | Grok 4.5 (high) | 93.1% |
| 8 | MiniMax-M3 | 92.9% |
| 9 | GPT-5.6 Sol (high) | 92.8% |
| 10 | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | 92.6% |
| 11 | GPT-5.6 Sol (medium) | 92.6% |
| 12 | GPT-5.5 (medium) | 92.6% |
| 13 | GPT-5.6 Terra (max) | 92.5% |
| 14 | Qwen3.7 Max | 92.3% |
| 15 | Gemini 3.5 Flash (high) | 92.2% |
| 16 | Gemini 3.5 Flash (medium) | 92.1% |
| 17 | Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | 92.0% |
| 18 | GPT-5.4 (xhigh) | 92.0% |
| 19 | GPT-5.3 Codex (xhigh) | 91.5% |
| 20 | Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | 91.4% |
| 21 | Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | 91.1% |
| 22 | GPT-5.6 Luna (max) | 91.1% |
| 23 | Kimi K2.6 | 91.1% |
| 24 | Grok 4.20 0309 v2 (Reasoning) | 91.1% |
| 25 | GPT-5.5 (low) | 91.0% |
| 26 | GPT-5.6 Terra (xhigh) | 90.8% |
| 27 | Gemini 3 Pro Preview (high) | 90.8% |
| 28 | DeepSeek V4 Pro (Reasoning, High Effort) | 90.5% |
| 29 | GPT-5.2 (xhigh) | 90.3% |
| 30 | Grok 4.3 (high) | 90.1% |
| 31 | Qwen3.7 Plus | 90.0% |
| 32 | GPT-5.2 Codex (xhigh) | 89.9% |
| 33 | Muse Spark 1.1 (xhigh) | 89.8% |
| 34 | GPT-5.6 Sol (low) | 89.8% |
| 35 | Gemini 3 Flash Preview (Reasoning) | 89.8% |
| 36 | Hy3 | 89.7% |
| 37 | GPT-5.6 Terra (high) | 89.6% |
| 38 | Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | 89.6% |
| 39 | Kimi K2.7 Code | 89.6% |
| 40 | GLM-5.2 (max) | 89.5% |
| 41 | GPT-5.6 Luna (xhigh) | 89.5% |
| 42 | Grok Build 0.1 0616 | 89.5% |
| 43 | DeepSeek V4 Flash (Reasoning, Max Effort) | 89.4% |
| 44 | Qwen3.5 397B A17B (Reasoning) | 89.3% |
| 45 | GPT-5.6 Luna (high) | 89.2% |
| 46 | Nex-N2-Pro | 89.2% |
| 47 | Grok 4.3 (medium) | 89.0% |
| 48 | DeepSeek V4 Pro (Reasoning, Max Effort) | 88.8% |
| 49 | Qwen3.6 Max Preview | 88.8% |
| 50 | Gemini 3 Pro Preview (low) | 88.7% |
Quality, pricing, and speed benchmarks via Artificial Analysis. Human preference Elo via LMArena. Refreshed daily at 06:00 UTC · last sync .
About the metrics
Intelligence Index
AA
Artificial Analysis composite of reasoning, knowledge, coding, math.
LMArena Elo
LMArena
Human preference Elo from chatbot arena head-to-head votes.
Coding Index
AA
AA composite — LiveCodeBench, SciCode, etc.
Math Index
AA
AA composite — MATH-500, AIME, etc.
MMLU-Pro
AA
Multi-domain knowledge & reasoning. Harder than MMLU.
GPQA Diamond
AA
Graduate-level science Q&A. Strong reasoning signal.
LiveCodeBench
AA
Competitive programming, contamination-resistant.
MATH-500
AA
Subset of the MATH benchmark.
AIME
AA
American Invitational Mathematics Examination — olympiad math.
Humanity's Last Exam
AA
Frontier reasoning benchmark; most models still score low.
SciCode
AA
Scientific coding tasks.