Frontier models, benchmarked continuously.
Quality, cost and speed for the leading AI models — refreshed on a live cycle · last updated 2026-08-23.
Intelligence, coding & agentic indexes
Composite indexes by Artificial Analysis · price per 1M tokens (input / output)
| # | Model | Intelligence | Coding | Agentic | $ / 1M in | $ / 1M out |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 5 (Adaptive Reasoning, Max Effort) | 63.1 | 78 | 59.2 | $5.50 | $27.50 |
| 2 | Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | 62.1 | 76.5 | 56.6 | $11.00 | $55.00 |
| 3 | GPT-5.6 Sol (max) | 60.9 | 77.4 | 57.8 | $5.50 | $33.00 |
| 4 | Grok 4.6 (high) | 60.9 | 76.8 | 58.7 | $2.00 | $6.00 |
| 5 | Kimi K3 (max) | 59.7 | 76.2 | 54.3 | $2.80 | $14.00 |
| 6 | GLM-5.3 (max) | 59.5 | 74.8 | 59.1 | $1.40 | $4.40 |
| 7 | Qwen3.8 Max | 58.1 | 71.8 | 58.4 | $2.00 | $6.00 |
| 8 | Qwen3.8 2.4T A95B | 57.7 | 71.9 | 57.1 | $2.00 | $6.00 |
| 9 | Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | 57.3 | 74.3 | 49.4 | $5.50 | $27.50 |
| 10 | Muse Spark 1.2 (xhigh) | 56.8 | 72.2 | 49.3 | $1.25 | $4.25 |
| 11 | GPT-5.6 Terra (max) | 56.6 | 76.7 | 50.2 | $2.20 | $13.20 |
| 12 | GPT-5.5 (xhigh) | 56.3 | 74.9 | 47.4 | $5.50 | $33.00 |
| 13 | Gemini 3.7 Flash (high) | 56 | 76.1 | 45.1 | $0.190 | $0.940 |
| 14 | Grok 4.5 (high) | 55.8 | 72.4 | 48.9 | $2.00 | $6.00 |
| 15 | Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | 55.3 | 71.5 | 49.7 | $2.20 | $11.00 |
| 16 | Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | 55 | 73.6 | 46.3 | $5.50 | $27.50 |
| 17 | Muse Spark 1.1 (xhigh) | 53.2 | 71.3 | 39.7 | $1.25 | $4.25 |
| 18 | DeepSeek V4 Pro 0813 (Reasoning, Max Effort) | 53.2 | 68.8 | 49.6 | $1.32 | $3.96 |
| 19 | GPT-5.4 (xhigh) | 53.1 | 71.1 | 44.2 | $2.75 | $16.50 |
| 20 | GLM-5.2 (max) | 52.6 | 68.8 | 45.7 | $0.500 | $3.15 |
| 21 | GPT-5.6 Luna (max) | 52.3 | 71.4 | 46.9 | $0.220 | $1.32 |
| 22 | Gemini 3.5 Flash (high) | 52 | 70.1 | 39.7 | $1.65 | $9.90 |
| 23 | Qwen3.8 27B | 52 | 68.1 | 50.9 | $0.400 | $3.00 |
| 24 | DeepSeek V4 Flash 0731 (Reasoning, Max Effort) | 51.8 | 69.1 | 48.4 | $0.050 | $0.100 |
| 25 | Gemini 3.6 Flash (high) | 51.6 | 69.2 | 40.5 | $0.830 | $4.12 |
| 26 | Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | 48.4 | 63 | 42.1 | $3.30 | $16.50 |
| 27 | Gemini 3.1 Pro Preview | 47.7 | 68.8 | 23 | $2.00 | $12.00 |
| 28 | Qwen3.7 Max | 46.7 | 66 | 30.9 | $1.48 | $4.42 |
| 29 | MiniMax-M3 | 45.4 | 58.6 | 36.1 | $0.280 | $1.10 |
| 30 | DeepSeek V4 Pro (Reasoning, Max Effort) | 45.3 | 59.4 | 37.8 | $0.410 | $0.810 |
Reasoning, agents & search — re-run on live models
GPQA Diamond
Graduate-level science reasoning
| # | Model | Score | $ / task |
|---|---|---|---|
| 1 | Gemini 3.1 Pro Preview | 94.3% | $0.204 |
| 2 | GPT-5.6 Sol Pro | 93.8% | $0.445 |
| 3 | GPT-5.5 | 93.3% | $0.303 |
| 4 | Gemini 3.5 Flash | 93.1% | $0.137 |
| 5 | Gemini 3.6 Flash | 92.4% | $0.090 |
| 6 | Gemini 3.7 Flash | 92.3% | $0.015 |
| 7 | GPT-5.6 Sol | 91.4% | $0.104 |
| 8 | Kimi K3 | 91.3% | $0.134 |
| 9 | GPT-5.6 Luna Pro | 90.7% | $0.047 |
| 10 | MiniMax M3 | 90.5% | $0.028 |
| 11 | GPT-5.4 | 90.4% | $0.153 |
| 12 | Nova Micro 1.0 | 89.7% | $0.206 |
| 13 | GPT-5.6 Terra | 89.6% | $0.021 |
| 14 | Claude Opus 4.7 | 88.7% | $0.225 |
| 15 | Claude Opus 4.8 | 88.6% | $0.189 |
No single model wins every benchmark.
The leader changes by task: one model tops reasoning, another tops coding, a third wins agentic search. That is exactly why MultipleChat exists — put the leaders side by side, let them draft, challenge and verify each other, and keep the best answer.
Quality, cost and speed together
A benchmark score without a price tag is marketing. Every leaderboard here pairs quality with cost per task or per million tokens.
Live & open
Refreshed continuously and downloadable as JSON. Our own quarterly comparison study lives at the MultipleChat AI Benchmark.
Attribution
Source: OpenRouter (openrouter.ai/rankings) incl. Artificial Analysis indexes, as of 2026-08-23T20:53:36Z. Licensed under CC BY 4.0.