Frontier Model Roster
Models compared on AA-Omniscience, intelligence, latency, speed, cost, and context. Data from Artificial Analysis.
| Model | Company | Context | AA-Omniscience Index↓ | AA-Agentic Index | AA-LCR | Visual Reasoning Index | Intelligence Index | Latency | Speed | Cost ($/M) | Intelligence / Cost | E2E / Cost | Ideal roles |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude Fable 5 (with fallback) | Anthropic | 1M | 40 | 53 | 70 | — | 60 | 148.11s | 66 | $7.70 | 7.79 | 0.001 | ★ PreferredExpert Consultant |
| Gemini 3.1 Pro Preview | 1M | 33 | 21 | 72.7 | 82 | 46 | 41.37s | 116 | $1.74 | 26.4 | 0.013 | Expert Consultant | |
| Claude Opus 4.8 (max) | Anthropic | 1M | 27 | 47 | 67.7 | — | 56 | 40.49s | 56 | $3.85 | 14.5 | 0.005 | Expert Consultant |
| Grok 4.5 (high) | SpaceXAI | 500K | 26 | 46 | 67.7 | 80 | 54 | 9.12s | 97 | $1.35 | 40.0 | 0.052 | ★ PreferredExpert ConsultantAll-rounder |
| Claude Opus 4.7 (max) | Anthropic | 1M | 26 | 44 | 70.3 | 79 | 54 | 24.13s | 50 | $3.85 | 14.0 | 0.008 | Expert Consultant |
| Gemini 3.5 Flash | 1M | 23 | 37 | 71 | 84 | 50 | 29.68s | 157 | $1.31 | 38.2 | 0.023 | Expert ConsultantAll-rounder | |
| GPT-5.6 Sol (max) | OpenAI | 1M | 22 | 54 | 73.7 | 83 | 59 | 145.61s | 54 | $4.35 | 13.6 | 0.001 | Expert Consultant |
| GPT-5.5 (xhigh) | OpenAI | 922K | 20 | 45 | 74.3 | 81 | 55 | 87.74s | 67 | $4.35 | 12.6 | 0.002 | Expert Consultant |
| Kimi K3 | Kimi | 1.1M | 18 | 50 | 74.7 | 81 | 57 | 1.99s | 62 | $2.31 | 24.7 | 0.010 | NEW★ PreferredExpert ConsultantSpeed |
| Muse Spark 1.1 (xhigh) | Meta | 1.1M | 18 | 38 | 63.3 | — | 51 | 1.32s | 117 | $0.78 | 65.4 | 0.057 | ★ PreferredExpert ConsultantHead/ExecutiveSpeedAll-rounder |
| Grok 4.3 (high) | SpaceXAI | 1M | 18 | 24 | 65 | 78 | 38 | 23.15s | 122 | $0.64 | 59.4 | 0.057 | ★ PreferredExpert ConsultantRoutineAll-rounder |
| Gemini 3 Pro Preview (high) | 1M | 16 | — | 70.7 | 80 | 40 | — | — | $1.74 | 23.0 | — | — | |
| Claude Sonnet 5 (max) | Anthropic | 1M | 15 | 47 | 70.7 | 77 | 53 | 198.73s | 78 | $1.54 | 34.4 | 0.003 | Expert Consultant |
| Grok 4.20 0309 v2 | SpaceXAI | 2M | 15 | — | 58 | 75 | 37 | 15.33s | 156 | $1.14 | 32.5 | 0.047 | All-rounder |
| Qwen3.7 Max | Alibaba | 1M | 14 | 31 | 69 | — | 46 | 2.50s | 202 | $1.43 | 32.2 | 0.041 | ★ Preferred |
| Claude Opus 4.6 (max) | Anthropic | 1M | 14 | — | 70.7 | 75 | 44 | 12.63s | 39 | $3.85 | 11.4 | 0.010 | — |
| Claude Opus 4.5 | Anthropic | 200K | 13 | — | 74 | 74 | 41 | 13.82s | 45 | $3.85 | 10.6 | 0.010 | — |
| Grok 4.20 0309 | SpaceXAI | 2M | 13 | — | 59 | 73 | 37 | — | — | $1.14 | 32.5 | — | — |
| Claude Sonnet 4.6 (max) | Anthropic | 1M | 12 | 41 | 70.7 | 73 | 47 | 116.28s | 52 | $2.31 | 20.3 | 0.003 | — |
| Gemini 3 Flash | 1M | 12 | — | 66.3 | 80 | 38 | 7.62s | 173 | $0.43 | 88.4 | 0.221 | RoutineAll-rounder | |
| GPT-5.3 Codex (xhigh) | OpenAI | 400K | 10 | — | 74 | 78 | 44 | 93.40s | 86 | $1.87 | 23.5 | 0.005 | — |
| Qwen3.6 Max Preview | Alibaba | 256K | 10 | — | 69.7 | — | 40 | 3.24s | 41 | $1.13 | 35.4 | 0.014 | — |
| GPT-5.5 Instant (May 2026) | OpenAI | 400K | 10 | 11 | 64 | — | 34 | — | — | $4.35 | 7.82 | — | — |
| Grok Build 0.1 0616 | SpaceXAI | 256K | 7 | 28 | 64.7 | 77 | 40 | 0.53s | 77 | $0.54 | 74.1 | 0.056 | SpeedRoutine |
| GPT-5.4 (xhigh) | OpenAI | 1.1M | 6 | 41 | 74 | 78 | 51 | 109.62s | 149 | $2.17 | 23.5 | 0.004 | Expert Consultant |
| Kimi K2.6 | Kimi | 256K | 6 | 30 | 69.7 | 79 | 44 | 2.54s | 46 | $0.70 | 62.9 | 0.013 | Routine |
| GPT-5.1 (high) | OpenAI | 272K | 6 | 21 | 75 | 75 | 37 | 33.31s | 102 | $1.34 | 27.6 | 0.020 | — |
| MiMo-V2-Pro | Xiaomi | 1M | 5 | — | 60.7 | — | 40 | — | — | — | — | — | — |
| GLM-5.2 (max) | Z AI | 1M | 4 | 43 | 71.3 | — | 51 | 1.55s | 159 | $0.90 | 56.7 | 0.064 | Expert ConsultantSpeed |
| MiMo-V2.5-Pro | Xiaomi | 1M | 4 | 29 | 73.3 | — | 42 | 2.74s | 57 | $0.18 | 233 | 0.120 | Routine |
| Muse Spark | Meta | 262K | 4 | 29 | 69.7 | 81 | 43 | — | — | — | — | — | — |
| Grok 4 | SpaceXAI | 256K | 4 | — | 68 | 69 | 33 | — | — | $7.70 | 4.29 | — | NEW |
| Qwen3.6 Plus | Alibaba | 1M | 3 | 28 | 69.7 | 78 | 40 | 2.55s | 53 | $0.43 | 93.0 | 0.020 | Routine |
| Inkling | Thinking Machines | 1M | 2 | 32 | 63.3 | 73 | 41 | 1.75s | 73 | $1.10 | 37.3 | 0.025 | NEWSpeed |
| GLM-5.1 | Z AI | 200K | 2 | 30 | 62.3 | — | 40 | 1.77s | 62 | $0.90 | 44.4 | 0.016 | Speed |
| GLM-5 | Z AI | 200K | 2 | — | 63.3 | — | 40 | 1.79s | 49 | $0.66 | 60.6 | 0.020 | SpeedRoutine |
| Qwen3.7 Plus | Alibaba | 1M | 2 | 21 | 65 | 80 | 39 | 2.82s | 53 | $0.27 | 144 | 0.074 | Routine |
| MiniMax-M3 | MiniMax | 1M | 1 | 35 | 74 | 79 | 44 | 1.94s | 84 | $0.22 | 200 | 0.143 | SpeedRoutine |
| MiniMax-M2.7 | MiniMax | 205K | 1 | 26 | 68.7 | — | 38 | 1.81s | 60 | $0.22 | 173 | 0.089 | SpeedRoutine |