Cost Comparison by Model

Per-token cost across providers for your task type and volume.

Showing per-model monthly cost across all 23 Prism catalog models for the Simple task. Cheapest model: Llama 3.1 8B (Groq) at $6.40/mo. Most expensive: Claude Opus 4.1 (legacy) at $3.5K/mo (539x cheaper-to-priciest spread).

RankModelProviderCost / reqMonthlyPrism pick
1Llama 3.1 8B (Groq)Groq$0.000064$6.40eco
2Llama 3.1 8B (Cerebras)Cerebras$0.000110$11
3Gemini 2.5 FlashGoogle$0.000150$15
4groq-gpt-oss-20bGroq$0.000150$15
5groq-gpt-oss-safeguard-20bGroq$0.000150$15
6Ministral 8BMistral$0.000165$17
7mistral-small-3-2Mistral$0.000170$17
8Llama 4 Scout (Groq)Groq$0.000190$19
9DeepSeek V4 FlashDeepSeek$0.000196$20
10fireworks-deepseek-v4-flashFireworks$0.000196$20
11gemini-2-5-flash-liteGoogle$0.000200$20
12GPT-4o miniOpenAI$0.000300$30
13gpt-oss 120B (Groq)Groq$0.000300$30
14groq-gpt-oss-120bGroq$0.000300$30
15mistral-small-4Mistral$0.000300$30
16Qwen3 32B (Groq)Groq$0.000409$41
17cerebras-gpt-oss-120bCerebras$0.000505$51
18CodestralMistral$0.000510$51
19fireworks-minimax-m2p7Fireworks$0.000600$60
20fireworks-minimax-m2p5Fireworks$0.000600$60
21gemini-3-1-flash-liteGoogle$0.000650$65
22Llama 3.3 70B (Groq)Groq$0.000709$71
23Qwen 235B (Cerebras)Cerebras$0.000840$84
24Mistral LargeMistral$0.000850$85
25mistral-large-3Mistral$0.000850$85
26magistral-smallMistral$0.000850$85
27devstral-2Mistral$0.000920$92
28gemini-2-5-flashGoogle$0.000990$99
29fireworks-kimi-k2p5Fireworks$0.001380$138
30groq-kimi-k2-instruct-0905Groq$0.001700$170
31GPT-5.4 MiniOpenAI$0.001950$195
32Kimi K2 (Fireworks)Fireworks$0.001960$196
33fireworks-kimi-k2p6Fireworks$0.001960$196
34Claude Haiku 4Anthropic$0.002300$230
35Claude Haiku 4.5Anthropic$0.002300$230
36DeepSeek V4 ProDeepSeek$0.002436$244
37fireworks-deepseek-v4-proFireworks$0.002436$244
38GLM 5.1 (Fireworks)Fireworks$0.002440$244
39cerebras-zai-glm-4-7Cerebras$0.002625$263
40Magistral MediumMistral$0.003100$310
41Mistral MediumMistral$0.003450$345
42mistral-medium-3-5Mistral$0.003450$345
43Gemini 3.5 FlashGoogle$0.003900$390
44Gemini 2.5 ProGoogle$0.004000$400
45gemini-2-5-proGoogle$0.004000$400
46GPT-4oOpenAI$0.005000$500
47gemini-3-1-pro-previewGoogle$0.005200$520
48Gemini 3.5 Pro (alias → 3.1 Pro)Google$0.005200$520
49GPT-5.4OpenAI$0.006500$650
50Claude Sonnet 4Anthropic$0.006900$690
51Claude Sonnet 4.7Anthropic$0.006900$690
52Claude Sonnet 4.6Anthropic$0.006900$690
53Claude Sonnet 4.5Anthropic$0.006900$690
54Claude Opus 4.7Anthropic$0.0115$1.1K
55Claude Opus 4.6Anthropic$0.0115$1.1K
56Claude Opus 4.5Anthropic$0.0115$1.1K
57GPT-5.5OpenAI$0.0130$1.3K
58Claude Opus 4Anthropic$0.0345$3.5Ksport
59Claude Opus 4.1 (legacy)Anthropic$0.0345$3.5K

Cost = (input_tokens × input_price + output_tokens × output_price) × volume. Prices are USD per 1M tokens, current Prism catalog (updated 2026-05-24). Highlighted models are what Prism's router picks for this task type in each mode (eco / balanced / sport). Cached requests cost $0 — model with the cache-hit-rate estimator for the full picture.

Frequently asked questions

Is Cost Comparison by Model really free?
Yes — fully free, no email gate, no signup, no sales call. All Prism tools live on this page and similar dedicated tool pages.
What data does Cost Comparison by Model use?
The calculations use current public provider pricing (Anthropic, OpenAI, Google) and Prism's published cache-economics math from production traffic. Nothing about your usage is captured — the calculator runs entirely in your browser.