LLM API Price Comparison 2026: The Same Models Across Platforms
Compare the same GPT, Claude, Gemini and DeepSeek API models on APIMaster, OpenRouter, Vercel AI Gateway and official APIs, with token rates and monthly costs.
Published
APIMaster's checked starting rates are about 93.3% below OpenRouter and Vercel AI Gateway for GPT-6.1 Sol, 87.1% below for Claude Opus 4.8, and 82.0% below for Claude Sonnet 4.6. For Gemini 3.8 Flash, the saving is about 63.4% against OpenRouter and 59.7% against Vercel and Google's current promotional price. OpenRouter's checked DeepSeek V4.1 Flash listing is cheaper than APIMaster. The useful comparison is the same model, timestamp and billing mode across platforms.
Checked November 10, 2026, 09:45 UTC. Rates are USD per 1 million input / output tokens, for ordinary paid, uncached, non-Batch inference. APIMaster includes only enabled public routes, excluding client-exclusive routes. Ranges pair the input and output prices of the lowest- and highest-input-priced routes. The live marketplace card controls current account prices and availability.
| Model ID | Official price / baseline | APIMaster current price | Active public routes | Discount vs official baseline | Model card |
|---|---|---|---|---|---|
gpt-6.1-sol |
$2 / $10 | $0.1345455 / $0.6727275 – $0.4755616 / $2.377808 | 13 | Up to ~93.3% off | Compare routes |
gpt-6-sol |
$2 / $10 | $0.1341243 / $0.6706215 – $0.4755616 / $2.377808 | 18 | Up to ~93.3% off | Compare routes |
claude-opus-4-8 |
$5 / $25 | $0.6445829 / $3.2229145 – $12 / $60 | 14 | Up to ~87.1% off | Compare routes |
claude-sonnet-4-6 |
$3 / $15 | $0.540468 / $2.70234 – $3 / $15 | 13 | Up to ~82.0% off | Compare routes |
claude-sonnet-5-5 |
$2 / $10 | $0.670392 / $3.35196 – $2 / $10 | 5 | Up to ~66.5% off | Compare routes |
gemini-3.8-flash |
$0.75 / $3.75 promotional | $0.3020126 / $1.5100627 – $1.5 / $7.5 | 7 | Up to ~59.7% off | Compare routes |
deepseek-flash |
$0.3 / $1.2 peak | $0.18 / $0.72 – $0.3 / $1.2 | 4 | Up to ~40.0% off peak | Compare routes |
Google's $0.75 / $3.75 promotion runs through December 31, 2026; its future $1.50 / $7.50 price is not today's comparison baseline. DeepSeek's official $0.30 / $1.20 is the peak-window rate at this snapshot; off-peak is $0.15 / $0.60. Some APIMaster routes, including the highest Opus 4.8 quote, cost more than official Standard. Choose a route rather than assuming every quote has the headline discount.
GPT Trial(GPT-6 Astra, GPT-6 Sol, GPT-5.6, GPT-5.5, gpt-5.6-luna, gpt-5.6-sol, gpt-5.6-terra, gpt-6.1-sol, gpt-6-astra, gpt-6-sol, and gpt-6-luna available)
Sign up and claim a $20 GPT trial
LLM API price comparison: APIMaster vs OpenRouter vs Vercel
Read across each row to compare platforms for the same model. A cheaper price on a different model is not counted as a platform saving. OpenRouter and Vercel figures are the public model-catalog quotes returned at the check, not a guarantee that every provider endpoint costs the same.
| Same model | APIMaster from | OpenRouter catalog | Vercel AI Gateway catalog | Official direct | APIMaster vs OpenRouter | APIMaster vs Vercel |
|---|---|---|---|---|---|---|
gpt-6.1-sol |
$0.1345455 / $0.6727275 | $2 / $10 | $2 / $10 | $2 / $10 | About 93.3% cheaper | About 93.3% cheaper |
gpt-6-sol |
$0.1341243 / $0.6706215 | $2 / $10 | $2 / $10 | $2 / $10 | About 93.3% cheaper | About 93.3% cheaper |
claude-opus-4-8 |
$0.6445829 / $3.2229145 | $5 / $25 | $5 / $25 | $5 / $25 | About 87.1% cheaper | About 87.1% cheaper |
claude-sonnet-4-6 |
$0.540468 / $2.70234 | $3 / $15 | $3 / $15 | $3 / $15 | About 82.0% cheaper | About 82.0% cheaper |
claude-sonnet-5-5 |
$0.670392 / $3.35196 | $2 / $10 | $2 / $10 | $2 / $10 | About 66.5% cheaper | About 66.5% cheaper |
gemini-3.8-flash |
$0.3020126 / $1.5100627 | $0.825 / $4.125 | $0.75 / $3.75 | $0.75 / $3.75 | About 63.4% cheaper | About 59.7% cheaper |
deepseek-flash |
$0.18 / $0.72 | $0.006 / $0.5 | $0.3 / $1.2 | $0.3 / $1.2 peak | OpenRouter is cheaper | About 40.0% cheaper |
The APIMaster IDs above are purchase links. OpenRouter and Vercel use openai/gpt-6.1-sol, openai/gpt-6-sol, anthropic/claude-opus-4.8, anthropic/claude-sonnet-4.6, anthropic/claude-sonnet-5.5 and google/gemini-3.8-flash. For DeepSeek, this comparison uses OpenRouter deepseek/deepseek-v4.1-flash-1017 and Vercel deepseek/deepseek-v4.1-flash; DeepSeek identifies its direct deepseek-flash alias as V4.1 Flash 1017. Check the selected provider's version and capabilities before migration.
Two findings change the buying decision. First, OpenRouter currently lists DeepSeek at $0.006 input / $0.50 output, below APIMaster's $0.18 / $0.72. Second, Gemini's checked OpenRouter quote is $0.825 / $4.125, while Vercel and Google's promotional Standard quote are $0.75 / $3.75. One generic “official price” column would hide these differences.
What does the same monthly workload cost on each platform?
Use 100 million uncached input tokens plus 20 million billed output tokens per month, spread over 1,000 calls of 100,000 input and 20,000 output tokens each. These calls stay within OpenAI's short-context price tier. This is quote arithmetic, not a paid performance benchmark or a promise that all models produce the same token counts.
Formula: monthly token cost = 100 × input price + 20 × output price.
| Same model | APIMaster from / month | OpenRouter / month | Vercel / month | Official direct / month | Difference vs OpenRouter |
|---|---|---|---|---|---|
gpt-6.1-sol |
$26.91 | $400.00 | $400.00 | $400.00 | Save $373.09 |
gpt-6-sol |
$26.82 | $400.00 | $400.00 | $400.00 | Save $373.18 |
claude-opus-4-8 |
$128.92 | $1000.00 | $1000.00 | $1000.00 | Save $871.08 |
claude-sonnet-4-6 |
$108.09 | $600.00 | $600.00 | $600.00 | Save $491.91 |
claude-sonnet-5-5 |
$134.08 | $400.00 | $400.00 | $400.00 | Save $265.92 |
gemini-3.8-flash |
$60.40 | $165.00 | $150.00 | $150.00 | Save $104.60 |
deepseek-flash |
$32.40 | $10.60 | $54.00 | $54.00 peak | OpenRouter saves $21.80 |
Keeping GPT-6.1 Sol and changing platform reduces this illustrative bill from $400 to about $26.91. Keeping Opus 4.8 reduces it from $1,000 to about $128.92. For DeepSeek, OpenRouter's listed rate gives $10.60 versus APIMaster's $32.40, so the direction reverses. Direct DeepSeek off-peak would be $27 for the same tokens, also below this APIMaster snapshot.
Prices exclude payment fees, taxes, tools, cache reads and writes, storage, retries and optional platform features. Include billed reasoning tokens in output. Neither a long aggregate monthly total nor the example's 1,000 calls creates a volume discount.
Platform fees, public discounts and credits
OpenRouter: its current pricing page lists a 5.5% credit-purchase fee on Standard and 8% on Business. The FAQ also states a $0.80 minimum for the card purchase fee and a 5% crypto fee. These are funding fees, not extra token rates. The tables exclude them so the comparison does not depend on payment method, plan or top-up size. Enterprise fees are contractual.
Vercel AI Gateway: its pricing page states zero token markup and no platform fee on tokens. The live catalog, including public promotions, is the rate source used here. Its free tier provides $5 per month for eligible models with limits; this is a separate offer, not a permanently lower paid token rate. Volume commitments may receive negotiated discounts, which are not assumed for a public pay-as-you-go comparison. Optional platform features can add charges.
APIMaster: compare the actual route price in your account before funding it. Public starting quotes do not assume a referral reward, negotiated private route or a credit bonus. A GPT trial promotion is separate from paid model pricing: read the current catalog and eligibility rather than assuming it funds Claude, Gemini or DeepSeek.
When can another platform still be cheaper?
Cache-heavy calls: GPT-6.1 Sol's direct Standard cache-read price is $0.10 per million tokens, below APIMaster's lowest uncached input quote of $0.1345455. Sonnet 4.6's direct cache reads are $0.30, below the $0.540468 uncached APIMaster quote. Compare cache writes, hit rates and storage too. The ordinary-request discount cannot be applied to a cache-heavy bill unchanged.
Batch, Flex and timing: OpenAI's Sol Standard $2 / $10 becomes $1 / $5 with Batch or Flex. Anthropic Batch has a 50% token discount. DeepSeek off-peak halves its official peak rates. These are separate billing conditions, not errors or hidden prices.
Long context and routing: OpenAI Sol requests above 272,000 input tokens have higher whole-request input and output rates. Regional processing, tools and provider selection can also change the price. OpenRouter and Vercel route among providers; APIMaster routes have different prices and supported features. A catalog minimum is a starting quote, not an endpoint-level service guarantee.
Compare the same accepted workload: model/version, endpoint, reasoning setting, tools, context, cache status, region and retries. Route-specific fingerprint and availability records help evaluate a route; they do not prove every response's identity or guarantee future uptime.
How to move the same model to APIMaster
Start with GPT-6.1 Sol's live routes, Opus 4.8 or Sonnet 4.6. Register, fund your APIMaster wallet and create an API key. Use https://apimaster.ai/v1 and the APIMaster model ID, not another platform's provider prefix.
curl "https://apimaster.ai/v1/responses" \
-H "Authorization: Bearer $APIMASTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"gpt-6.1-sol","input":"Summarize the following support ticket in three sentences.","max_output_tokens":256}'
This is a minimal example. Keep your real prompts and scoring rubric unchanged, check the route's endpoint and tool support, and compare returned usage and the first small bill before moving production traffic. OpenAI requires Responses for GPT-6.1 Sol tool calling. An APIMaster top-up funds APIMaster, not an OpenAI, Anthropic, OpenRouter or Vercel account.
FAQ
Is APIMaster always cheaper than OpenRouter?
No. The checked GPT and Claude starting rates are substantially lower, but OpenRouter's DeepSeek V4.1 Flash quote is lower than APIMaster's. Compare the exact model and provider at the time of purchase.
Is Vercel AI Gateway more expensive because it adds a markup?
Its pricing page says zero token markup. The lower APIMaster quotes shown here come from its available route prices, not from an assumed Vercel markup.
Why is Gemini's discount smaller than a marketplace reference suggests?
Google currently charges $0.75 / $3.75 through December 31, 2026, versus the later $1.50 / $7.50 price. This comparison uses today's promotional price, giving about 59.7% savings at the checked APIMaster starting quote.
Does this compare LLM models or API platforms?
Platforms. Every row holds the model fixed. For broader AI API price comparison, see the separate image API provider comparison and video API provider comparison.
Are the displayed monthly costs measured bills?
No. They are reproducible estimates from the checked per-token quotes. Actual costs depend on route, cached tokens, reasoning output, tools, retries and your workload.
Sources and model-specific guides
Price snapshots checked November 10, 2026:
- APIMaster live marketplace; public per-model pricing endpoint
/api/marketplace-data?model=<model-id>. - OpenRouter live model catalog, platform pricing and fee FAQ.
- Vercel AI Gateway model catalog, pricing and discounts.
- OpenAI API pricing and GPT-6.1 Sol.
- Anthropic pricing, Google Gemini pricing and DeepSeek pricing/version mapping.
For individual routes and setup: Cheap GPT-6.1 Sol · Cheap Opus 4.8 · Cheap Sonnet 4.6 · OpenRouter vs APIMaster.
Compare your existing model's live price → register → create an API key → use https://apimaster.ai/v1 with gpt-6.1-sol.