Best AI Models for Product Managers
Frontier models ranked on the Artificial Analysis Intelligence Index, which every listed model has. The PM Index is shown only when the Agentic Index and the Coding Index are both present. The slider leans toward a lower list price, not the cost of a task.
Updated October 29, 2026
Model recommender
Which AI model should I use?
Pick your task and budget for a data-backed recommendation.
Recommended for writing prds
Gemini 4 Carbon (High)
See the full ranking for Writing PRDs →Intelligence 59.4 · list $8.00/1M
Value frontier chart
Higher is the Intelligence Index. Left is a lower list price. Ringed models are not beaten on both intelligence and list price. Hover a point, or a lab below.
- Anthropic
- OpenAI
- Kimi
- Meta
- xAI
- Alibaba
- DeepSeek
- Z AI
- Xiaomi
- MiniMax
- StepFun
Movers · last 7 days
▲ Climbing
- Command-R+ (Apr '24)▲70 · #36
- Exaone 4.0 1.2B (Non-reasoning)▲65 · #42
- ERNIE 4.5 300B A47B▲53 · #40
- EXAONE 4.0 32B (Non-reasoning)▲46 · #43
- Jamba 1.5 Mini▲45 · #63
▼ Falling
- Kimi K3 (Low)▼50 · #71
- Solar Mini 4▼44 · #97
- Step 3.7 Flash▼44 · #108
- Step 3.5 Flash 2603▼42 · #107
- Step 3.5 Flash▼40 · #106
AI model rankings
Highest intelligenceLower list price
| # | Model | Intelligence | PM Index | List $/1M | Adjusted |
|---|---|---|---|---|---|
| 1 | Gemini 4 Carbon (High) Google | 59.4 | — | $8.00 | 10.0 |
| 2 | Claude Opus 5.5 (Max) Anthropic | 58.4 | — | $8.00 | 9.8 |
| 3 | GPT-6 Astra (Max) OpenAI | 56.9 | — | $20.00 | 9.6 |
| 4 | GPT-6.1 Sol (Max) OpenAI | 54.5 | — | $4.00 | 9.2 |
| 5 | Claude Sonnet 5.5 (Max) Anthropic | 54.2 | — | $4.00 | 9.1 |
| 6 | GPT-6.1 Terra (Max) OpenAI | 51.0 | — | $2.00 | 8.6 |
| 7 | Gemini 4 Argon (High) Google | 50.4 | — | $8.00 | 8.5 |
| 8 | Claude Fable 5.1 (Xhigh) Anthropic | 49.7 | — | $20.00 | 8.4 |
| 9 | Muse Wonder Full (Max) Meta | 49.7 | — | — | 8.4 |
| 10 | Kimi K3.1 Pro Kimi | 46.0 | — | $4.00 | 7.7 |
| 11 | Muse Wonder (Max) Meta | 43.3 | — | $8.00 | 7.3 |
| 12 | Muse Spark 1.3 (Max) Meta | 36.4 | — | $2.00 | 6.1 |
| 13 | Claude Haiku 5.5 (Max) Anthropic | 36.2 | — | $0.20 | 6.1 |
| 14 | Grok 4.7 (High) SpaceXAI | 35.0 | — | $3.00 | 5.9 |
| 15 | Qwen3.8 Max (0902) Alibaba | 33.6 | — | $3.00 | 5.7 |
| 16 | DeepSeek V4.1 Flash 1017 (Max) DeepSeek | 32.5 | — | $0.52 | 5.5 |
| 17 | GLM-5.3 (Max) Z AI | 31.7 | — | $2.15 | 5.3 |
| 18 | Gemini 3.8 Flash (High) Google | 30.5 | — | $1.50 | 5.1 |
| 19 | MiMo-V2.6-Pro Xiaomi | 29.9 | — | $0.54 | 5.0 |
| 20 | MiniMax-M3.1-Flash MiniMax | 28.9 | — | $0.26 | 4.9 |
| 21 | Step 5 Preview StepFun | 28.5 | — | $1.43 | 4.8 |
| 22 | GPT-6 Luna (Max) OpenAI | 27.8 | — | $0.20 | 4.7 |
| 23 | MiMo-V2.6-Flash Xiaomi | 26.7 | — | $0.18 | 4.5 |
| 24 | Mistral Large 4 (High) Mistral | 26.0 | — | $1.05 | 4.4 |
| 25 | Mistral Large 4 Preview Mistral | 25.3 | — | $2.06 | 4.3 |
| 26 | Muse Spark Open (Xhigh) Meta | 24.7 | — | $2.00 | 4.2 |
| 27 | DeepSeek V4 Pro 0813 (Max) DeepSeek | 23.1 | — | $1.98 | 3.9 |
| 28 | Qwen3.8 27B (Xhigh) Alibaba | 20.1 | — | $1.13 | 3.4 |
| 29 | MiniMax-M3 MiniMax | 18.7 | — | $0.52 | 3.1 |
| 30 | Nemotron 3 Ultra 550B A55B (Reasoning) NVIDIA | 11.8 | — | $1.13 | 2.0 |
| 31 | Gemini 3.5 Flash-Lite Google | 11.7 | — | $0.85 | 2.0 |
| 32 | Muse Glimmer (High) Meta | 6.3 | — | $0.52 | 1.1 |
| 33 | Command A+ Cohere | 6.2 | — | $0.00 | 1.0 |
| 34 | NVIDIA Nemotron 3 Nano 4B NVIDIA | 1.8 | — | — | 0.3 |
| 35 | Claude 3.5 Haiku Anthropic | — | — | — | 0.0 |
| 36 | Command-R+ (Apr '24) Cohere | — | — | — | 0.0 |
| 37 | DeepSeek R1 (Jan '25) DeepSeek | — | — | $2.50 | 0.0 |
| 38 | DeepSeek V3.1 Terminus (Reasoning) DeepSeek | — | — | $1.91 | 0.0 |
| 39 | DeepSeek V3.2 Exp (Non-reasoning) DeepSeek | — | — | $0.32 | 0.0 |
| 40 | ERNIE 4.5 300B A47B Baidu | — | — | — | 0.0 |
| 41 | ERNIE 5.0 Thinking Preview Baidu | — | — | — | 0.0 |
| 42 | Exaone 4.0 1.2B (Non-reasoning) LG AI Research | — | — | — | 0.0 |
| 43 | EXAONE 4.0 32B (Non-reasoning) LG AI Research | — | — | — | 0.0 |
| 44 | EXAONE 4.5 33B (Non-reasoning) LG AI Research | — | — | — | 0.0 |
| 45 | Gemini 3.5 Flash (High) Google | — | — | $3.38 | 0.0 |
| 46 | GLM-4.5 (Reasoning) Z AI | — | — | — | 0.0 |
| 47 | GLM-4.7-Flash (Reasoning) Z AI | — | — | $0.15 | 0.0 |
| 48 | GLM-5-Turbo Z AI | — | — | — | 0.0 |
| 49 | GLM-5.3-Flash Z AI | — | — | $0.24 | 0.0 |
| 50 | GPT-6 Sol (Low) OpenAI | — | — | $4.00 | 0.0 |
| 51 | Granite 3.3 8B (Non-reasoning) IBM | — | — | — | 0.0 |
| 52 | Granite 4.0 350M IBM | — | — | — | 0.0 |
| 53 | Granite 4.0 Micro IBM | — | — | — | 0.0 |
| 54 | Granite 4.2 30B IBM | — | — | — | 0.0 |
| 55 | Granite 4.2 3B IBM | — | — | $0.05 | 0.0 |
| 56 | Grok 3 SpaceXAI | — | — | $6.00 | 0.0 |
| 57 | Grok 4.1 Fast (Reasoning) SpaceXAI | — | — | — | 0.0 |
| 58 | Grok 4.20 0309 (Reasoning) SpaceXAI | — | — | $3.00 | 0.0 |
| 59 | Grok Build 0.1 0616 SpaceXAI | — | — | $1.25 | 0.0 |
| 60 | Hy3 Tencent | — | — | $0.24 | 0.0 |
| 61 | Hy3-preview (Reasoning) Tencent | — | — | $0.10 | 0.0 |
| 62 | Jamba 1.5 Large AI21 Labs | — | — | $3.50 | 0.0 |
| 63 | Jamba 1.5 Mini AI21 Labs | — | — | $0.25 | 0.0 |
| 64 | Jamba 1.6 Large AI21 Labs | — | — | — | 0.0 |
| 65 | Jamba 1.7 Large AI21 Labs | — | — | — | 0.0 |
| 66 | Jamba 1.7 Mini AI21 Labs | — | — | — | 0.0 |
| 67 | K-EXAONE (Reasoning) LG AI Research | — | — | — | 0.0 |
| 68 | K-EXAONE 2.0 0803 LG AI Research | — | — | — | 0.0 |
| 69 | Kimi K2 Kimi | — | — | $1.00 | 0.0 |
| 70 | Kimi K2.6 (Reasoning) Kimi | — | — | $1.71 | 0.0 |
| 71 | Kimi K3 (Low) Kimi | — | — | $6.00 | 0.0 |
| 72 | Kimi Linear 48B A3B Instruct Kimi | — | — | — | 0.0 |
| 73 | Llama 3.3 Nemotron Super 49B v1 (Reasoning) NVIDIA | — | — | — | 0.0 |
| 74 | Llama Nemotron Super 49B v1.5 (Non-reasoning) NVIDIA | — | — | — | 0.0 |
| 75 | Magistral Small 1 Mistral | — | — | — | 0.0 |
| 76 | MiMo-V2-Flash (Feb 2026) Xiaomi | — | — | — | 0.0 |
| 77 | MiMo-V2.5 Xiaomi | — | — | $0.17 | 0.0 |
| 78 | MiMo-V2.5-Pro (Reasoning) Xiaomi | — | — | $0.56 | 0.0 |
| 79 | MiniMax M1 40k MiniMax | — | — | — | 0.0 |
| 80 | MiniMax-M2.1 MiniMax | — | — | $0.52 | 0.0 |
| 81 | MiniMax-M2.5 MiniMax | — | — | $0.52 | 0.0 |
| 82 | Mistral Medium 3.1 Mistral | — | — | — | 0.0 |
| 83 | Mistral Small 4 (Non-reasoning) Mistral | — | — | $0.26 | 0.0 |
| 84 | Nova 2.0 Lite (High) Amazon | — | — | $0.85 | 0.0 |
| 85 | Nova 2.0 Pro Preview (Medium) Amazon | — | — | $3.44 | 0.0 |
| 86 | Nova Lite Amazon | — | — | $0.10 | 0.0 |
| 87 | Nova Micro Amazon | — | — | $0.06 | 0.0 |
| 88 | Nova Pro Amazon | — | — | $1.40 | 0.0 |
| 89 | NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning) NVIDIA | — | — | $0.09 | 0.0 |
| 90 | Phi-3 Mini Instruct 3.8B Microsoft | — | — | — | 0.0 |
| 91 | Phi-4 Microsoft | — | — | $0.22 | 0.0 |
| 92 | Phi-4 Mini Instruct Microsoft | — | — | $0.13 | 0.0 |
| 93 | Qwen3 0.6B (Non-reasoning) Alibaba | — | — | — | 0.0 |
| 94 | Qwen3.5 122B A10B (Non-reasoning) Alibaba | — | — | $1.10 | 0.0 |
| 95 | Qwen3.5 2B (Non-reasoning) Alibaba | — | — | — | 0.0 |
| 96 | R1 1776 Perplexity | — | — | — | 0.0 |
| 97 | Solar Mini 4 Upstage | — | — | $0.18 | 0.0 |
| 98 | Solar Open 100B (Reasoning) Upstage | — | — | — | 0.0 |
| 99 | Solar Pro 2 (Non-reasoning) Upstage | — | — | — | 0.0 |
| 100 | Solar Pro 2 (Preview) (Non-reasoning) Upstage | — | — | — | 0.0 |
| 101 | Solar Pro 4 Upstage | — | — | $0.52 | 0.0 |
| 102 | Sonar Perplexity | — | — | — | 0.0 |
| 103 | Sonar Pro Perplexity | — | — | — | 0.0 |
| 104 | Sonar Reasoning Perplexity | — | — | — | 0.0 |
| 105 | Sonar Reasoning Pro Perplexity | — | — | — | 0.0 |
| 106 | Step 3.5 Flash StepFun | — | — | $0.15 | 0.0 |
| 107 | Step 3.5 Flash 2603 StepFun | — | — | $0.15 | 0.0 |
| 108 | Step 3.7 Flash StepFun | — | — | $0.44 | 0.0 |
| 109 | Tiny Aya Global Cohere | — | — | $0.00 | 0.0 |
Get the monthly PM AI rankings
New frontier models ship constantly. Get the updated Intelligence Index rankings, and which model to use for which PM task, once a month. No spam, unsubscribe anytime.
Best AI model by PM use case
The right model depends on the job. These rankings reweight the benchmarks for specific product-management tasks.
How the ranking works
The board is ranked on the Artificial Analysis Intelligence Index, which every listed model has. The PM Index is 70% the Artificial Analysis Intelligence Index, 20% the Agentic Index, and 10% the Coding Index, on the same 0 to 100 scale. It is shown only when the Agentic Index and the Coding Index are both present. List price is the Artificial Analysis blended rate, 3 parts input to 1 part output, per 1M tokens. That price is not the cost of a task.