Updated Oct 18, 2026
What AI really costs. Per task, not per token.
The price of a chat answer, a PDF summary or a code review in 336 models, measured on real prompts.
336models priced
14everyday tasks
1,331provider prices
12models measured
Cost per task, today
Chat & writingDocuments & dataCode & agentsMedia1K image →4K image →1 min video →1 min voice →1 h transcription →
| # | Model | Cost per task | Quality | Tokens in / out | Basis | |
|---|---|---|---|---|---|---|
| 1 |  GLM 5.3 FlashZ.ai | 0.04¢$0.36 / 1,000 | 1474 | 48 / 1,782 | measured | |
| 2 |  GPT-6 LunaOpenAI | 0.06¢$0.59 / 1,000 | 1443 | 45 / 1,179 | measured | |
| 3 |  Claude Haiku 5.5Anthropic | 0.07¢$0.66 / 1,000 | - | 61 / 1,299 | measured | |
| 4 |  Qwen3.8 FlashAlibaba Qwen | 0.08¢$0.76 / 1,000 | - | 88 / 1,598 | measured | |
| 5 |  Mistral Small 4Mistral | 0.08¢$0.83 / 1,000 | - | 40 / 1,377 | estimated | |
| 6 |  MiMo-V2.6-ProXiaomi | 0.09¢$0.90 / 1,000 | 1479 | 206 / 933 | estimated | |
| 7 |  DeepSeek V4.1 FlashDeepSeek | 0.09¢$0.91 / 1,000 | 1475 | 65 / 6,320 | measured | |
| 8 |  Llama 4 MaverickMeta | 0.11¢$1.06 / 1,000 | 1327 | 87 / 1,592 | estimated | |
| 9 |  MiniMax M3MiniMax | 0.12¢$1.18 / 1,000 | 1441 | 205 / 929 | measured | |
| 10 |  Command A+Cohere | 0.15¢$1.46 / 1,000 | - | 205 / 930 | estimated | |
| 11 | GPT-5.4 NanoOpenAI | 0.15¢$1.48 / 1,000 | 1401 | 45 / 1,179 | estimated | |
| 12 |  Gemini 3.1 Flash LiteGoogle | 0.25¢$2.46 / 1,000 | - | 35 / 1,631 | estimated | |
| 13 | Mistral Large 4Mistral | 0.29¢$2.91 / 1,000 | 1432 | 40 / 1,377 | measured | |
| 14 |  Nemotron 3 UltraNVIDIA | 0.33¢$3.31 / 1,000 | 1426 | 39 / 1,370 | estimated | |
| 15 | Gemini 3.5 Flash LiteGoogle | 0.41¢$4.09 / 1,000 | 1455 | 35 / 1,631 | estimated | |
| 16 |  Grok 4.3xAI | 0.45¢$4.53 / 1,000 | 1442 | 1,280 / 1,170 | estimated | |
| 17 | GPT-5.4 MiniOpenAI | 0.53¢$5.34 / 1,000 | 1447 | 45 / 1,179 | estimated | |
| 18 | GPT-6.1 SolOpenAI | 0.67¢$6.72 / 1,000 | 1484 | 45 / 663 | measured | |
| 19 | Gemini 3.8 FlashGoogle | 0.68¢$6.76 / 1,000 | 1496 | 35 / 1,631 | measured | |
| 20 | GLM 5.3Z.ai | 0.86¢$8.56 / 1,000 | 1478 | 48 / 1,782 | estimated | |
| 21 | Grok 4.7xAI | 0.96¢$9.58 / 1,000 | 1444 | 1,280 / 1,170 | measured | |
| 22 | DeepSeek V4 Pro 0813DeepSeek | 1.3¢$12.56 / 1,000 | - | 65 / 6,320 | estimated | |
| 23 | Claude Sonnet 5.5Anthropic | 1.3¢$12.68 / 1,000 | 1475 | 61 / 1,256 | measured | |
| 24 |  Kimi K3Moonshot AI | 1.6¢$16.32 / 1,000 | 1487 | 39 / 1,366 | estimated | |
| 25 | Qwen3.8 Max PrimeAlibaba Qwen | 2.0¢$19.53 / 1,000 | - | 88 / 1,598 | estimated | |
| 26 | Gemini 3.1 Pro PreviewGoogle | 2.1¢$21.39 / 1,000 | 1487 | 35 / 1,777 | measured | |
| 27 | Claude Opus 5.5Anthropic | 2.5¢$25.36 / 1,000 | 1508 | 61 / 1,256 | estimated | |
| 28 | GPT-6 AstraOpenAI | 3.4¢$33.60 / 1,000 | 1477 | 45 / 663 | estimated | |
| 29 | Claude Fable 5.1Anthropic | 6.3¢$63.41 / 1,000 | 1502 | 61 / 1,256 | estimated |
29 popular models. Cost = tokens used on this task x the model's current list price.
Full ranking
Cheapest popular model for every task
Answer a chat question0.04¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.09¢Reply to a customer email<0.01¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.06¢Write a 1,500-word blog post0.04¢DeepSeek V4.1 Flashtop-rated pick: MiMo-V2.6-Pro 0.20¢Write 10 product descriptions0.04¢DeepSeek V4.1 Flashtop-rated pick: MiMo-V2.6-Pro 0.16¢Translate 1,000 words0.08¢Qwen3.8 Flashtop-rated pick: MiMo-V2.6-Pro 0.57¢Proofread one page (500 words)0.02¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.13¢Summarize a 20-page PDF0.10¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.63¢Extract data from an invoice0.02¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.10¢Analyze a 200-row spreadsheet0.19¢GPT-6 Lunatop-rated pick: MiMo-V2.6-Pro 1.6¢Answer from 5 retrieved documents (RAG)0.05¢GPT-6 Lunatop-rated pick: MiMo-V2.6-Pro 0.28¢Classify 100 support tickets0.06¢GPT-6 Lunatop-rated pick: MiMo-V2.6-Pro 0.23¢Review 500 lines of code0.26¢DeepSeek V4.1 Flashtop-rated pick: GPT-6.1 Sol 4.9¢Write a SQL query0.03¢DeepSeek V4.1 Flashtop-rated pick: GPT-6.1 Sol 0.48¢Run a 50-step coding agent2.4¢DeepSeek V4.1 Flashtop-rated pick: GPT-6.1 Sol 53¢Generate one 1K image$0.0059GPT Image 2.5 (Sunburst / Flare)via OpenAIGenerate one 4K image$0.113Nano Banana 2.1via GoogleOne minute of 1080p AI video$4.80Veo 3.1 Litevia GoogleOne minute of AI voice (TTS)$0.0036WaveNet voicesvia Google CloudTranscribe one hour of audio$0.040
Whisper Large v3 Turbovia GroqYour whole month$ / moMonthly AI bill calculatorevery model, your volumes →
GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.09¢Reply to a customer email<0.01¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.06¢Write a 1,500-word blog post0.04¢DeepSeek V4.1 Flashtop-rated pick: MiMo-V2.6-Pro 0.20¢Write 10 product descriptions0.04¢DeepSeek V4.1 Flashtop-rated pick: MiMo-V2.6-Pro 0.16¢Translate 1,000 words0.08¢Qwen3.8 Flashtop-rated pick: MiMo-V2.6-Pro 0.57¢Proofread one page (500 words)0.02¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.13¢Summarize a 20-page PDF0.10¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.63¢Extract data from an invoice0.02¢GLM 5.3 Flashtop-rated pick: MiMo-V2.6-Pro 0.10¢Analyze a 200-row spreadsheet0.19¢GPT-6 Lunatop-rated pick: MiMo-V2.6-Pro 1.6¢Answer from 5 retrieved documents (RAG)0.05¢GPT-6 Lunatop-rated pick: MiMo-V2.6-Pro 0.28¢Classify 100 support tickets0.06¢GPT-6 Lunatop-rated pick: MiMo-V2.6-Pro 0.23¢Review 500 lines of code0.26¢DeepSeek V4.1 Flashtop-rated pick: GPT-6.1 Sol 4.9¢Write a SQL query0.03¢DeepSeek V4.1 Flashtop-rated pick: GPT-6.1 Sol 0.48¢Run a 50-step coding agent2.4¢DeepSeek V4.1 Flashtop-rated pick: GPT-6.1 Sol 53¢Generate one 1K image$0.0059GPT Image 2.5 (Sunburst / Flare)via OpenAIGenerate one 4K image$0.113Nano Banana 2.1via GoogleOne minute of 1080p AI video$4.80Veo 3.1 Litevia GoogleOne minute of AI voice (TTS)$0.0036WaveNet voicesvia Google CloudTranscribe one hour of audio$0.040Whisper Large v3 Turbovia GroqYour whole month$ / moMonthly AI bill calculatorevery model, your volumes →Value for money: chat answer
Each dot is a model: cost of one chat answer (left = cheaper) against its quality score (up = better). Green dots are the value frontier - no other model is both cheaper and better.