Live prices · updated Nov 7, 2026

Every AI model.
Every provider.
The real price.

Same model, different bill: gpt-oss-120b costs $26 a month at W&B Inference (CoreWeave) and $180 at Cloudflare Workers AI for the same workload. Search 1,263 AI models across 58 providers and see who sells each one cheapest.

Free, no sign-up. No provider pays for its ranking.

Start with what you want to do

What are you building?

One task. Relevant prices. No mixed leaderboard.

LLM API pricing

Same model, different price.

274 LLMs that at least two providers sell, ranked by who sells each one cheapest.

274 models, showing 10

Sort and filter in the column headings. Prices blend 3 input : 1 output tokens; open a model for your own volume, or tick two to compare them. GPQA: graduate-level science questions answered correctly in Epoch AI's runs (CC BY 4.0).

Developer
gpt-oss-120bOpenAI196.9×W&B Inference (CoreWeave)$0.03$0.1776%131K
GLM-5.2Z.ai142.6×DeepInfra$0.75$2.4092%1M
Llama-3.3-70B-InstructMeta135.1×Nebius Token Factory$0.13$0.4047%131K
GLM-5.3Z.ai121.5×DeepInfra$0.90$4.0091%1M
GLM-5.3-FlashZ.ai11same priceZ.ai$0.15$0.5090%1M
DeepSeek V4 Flash 0731DeepSeek106.3×DeepInfra$0.08$0.1891%1M
gpt-oss-20bOpenAI104.1×W&B Inference (CoreWeave)$0.03$0.1361%131K
Kimi K3Moonshot AI101.1×SiliconFlow$2.70$13.5093%1M
Llama-3.1-8B-InstructMeta108.0×Novita AI$0.02$0.0527%131K
DeepSeek-V4.1-FlashDeepSeek92.5×DeepInfra$0.14$0.4290%1M

Standard list prices from each provider's pricing page, checked daily. Ranked by price only; no provider pays for its position. How we collect and rank prices

Quick estimate

Price your monthly video workflow

Start with three inputs. We will carry them into the full comparison and start with a clearly labeled retry assumption you can change.

Sources checked Oct 27, 2026 to Nov 3, 2026

Starts with 3 attempts per usable clip. Change it in the full calculator.

Latest

What changed, and what we think about it

No-BS Take · Nov 2, 2026

DeepSeek V4.1 Flash Is Half Price Off-Peak, but Most Hosts Charge the Peak Rate

DeepSeek charges $0.15 / $0.60 for V4.1 Flash outside its peak hours, which cover only 35 of 168 hours a week. Together, Novita and Fireworks list the peak rate of $0.30 / $1.20 around the clock. DeepInfra is cheaper still ($0.14 / $0.42 during a promotion, $0.20 / $0.60 regular). If you pay $0.30 / $1.20 today, you can halve the bill.

Recent price moves