By the toolist.ai editors. Researched from the providers' own pricing pages, not hands-on tested by toolist.ai. Prices checked November 3, 2026; we re-check them every week.

AI model API prices and comparisons

Of the newest models from OpenAI, Google, Meta, Anthropic and Z.ai, Claude Haiku 5.5 and GPT-6 Luna are the cheapest at $0.10 per million input tokens and $0.50 per million output tokens. GPT-6 Astra and Claude Fable 5.1 are the most expensive at $10 and $50. The table shows what each one costs for the same support chatbot workload, so you can compare a month's bill, not just a token price.

API price table

ModelInputCached inputOutputContext windowReleasedSupport chatbot, a month
GPT-6.1 Terra
OpenAI
$1 $0.05 $5 1.05M 2026-10-26 $500
Gemini 4 Argon
Google
$2 $0.10 $10 1.05M 2026-10-13 $1,000
Muse Wonder
Meta
$4 $0.20 $20 1.05M 2026-10-12 $2,000
Claude Haiku 5.5
Anthropic
$0.10 $0.01 $0.50 1M 2026-10-07 $50.00
GPT-6.1 Sol
OpenAI
$2 $0.10 $10 1.05M 2026-09-29 $1,000
Claude Sonnet 5.5
Anthropic
$2 $0.10 $10 1M 2026-09-28 $1,000
Claude Opus 5.5
Anthropic
$4 $0.20 $20 1M 2026-09-22 $2,000
GPT-6 Luna
OpenAI
$0.10 $0.01 $0.50 1.05M 2026-09-22 $50.00
GPT-6 Sol
OpenAI
$2 $0.20 $10 1.05M 2026-09-22 $1,000
GPT-6 Astra
OpenAI
$10 $1 $50 1.05M 2026-09-03 $5,000
Claude Fable 5.1
Anthropic
$10 $0.25 $50 1M 2026-09-01 $5,000
GLM-5.3
Z.ai
$1.40 $0.26 $4.40 1M 2026-08-18 $596
Claude Opus 5
Anthropic
$5 $0.50 $25 1M 2026-07-24 $2,500

USD per million tokens at standard rates, newest model first. Support chatbot: 100,000 replies, each 3,000 input and 400 output tokens, no caching. Each model's page lists long-context, batch and fast-mode prices.

Announced but not on sale, so not in this table: Gemini 4 Carbon.

Price your own workload

ModelCost a month
Claude Haiku 5.5 $7.50
GPT-6 Luna $7.50
GPT-6.1 Terra $75.00
GLM-5.3 $92.00
Claude Sonnet 5.5 $150
Gemini 4 Argon $150
GPT-6 Sol $150
GPT-6.1 Sol $150
Claude Opus 5.5 $300
Muse Wonder $300
Claude Opus 5 $375
Claude Fable 5.1 $750
GPT-6 Astra $750

Cheapest first. Standard API prices, no batch or priority discounts. Cached input is billed at the cache-read price; cache-write charges are not included. Reasoning models bill their thinking as output tokens.

Head-to-head comparisons

The models

Other workloads

ModelDocument Q&ACoding agent
GPT-6.1 Terra$63.50$93.50
Gemini 4 Argon$127$187
Muse Wonder$254$374
Claude Haiku 5.5$7.70$10.70
GPT-6.1 Sol$127$187
Claude Sonnet 5.5$127$187
Claude Opus 5.5$254$374
GPT-6 Luna$7.70$10.70
GPT-6 Sol$154$214
GPT-6 Astra$770$1,070
Claude Fable 5.1$568$868
GLM-5.3$130$156
Claude Opus 5$385$535

Document Q&A: 5,000 questions over the same 60,000-token document set (90% cache hits), 800 output tokens each (300M in, 4M out). Coding agent: 200 tasks, each 1.5 million input tokens (90% cache hits) and 50,000 output tokens (300M in, 10M out). Cached input is billed at the cache-read price; cache-write charges are not included.

How we gather and check these numbers: our method.

Sources