AI Model Finder
Find the cheapest AI model for your workload
Compare every major LLM side-by-side. Sorted by price. Filter by context, modality, or compliance.
Pricing verified: 2026-07-28
163 models, 16 providers
🔄 Refreshed weekly
What this calculator does
Filter the full catalog of 150+ AI models by your role, workload, capability needs, and price ceiling — output is a ranked shortlist you can ship to other calculators.
Why use it
- Stop reading vendor blog posts — every model is in one table, freshness-stamped
- Filter by what you actually need (vision, long-context, tier-1 only) instead of skimming marketing pages
- Price slider + min-context slider narrow 150+ models to the 5-10 worth comparing
- Compare checkbox lets you hold 2-4 candidates side-by-side before exporting to Cost Calculator
Who uses this:
Vibe Coder
High
Quick way to discover the new model that just launched without reading 10 launch posts
Small Business
High
Filter to Tier-1 + budget cap quickly when you need to swap a model and don't have time to research
Enterprise
Medium
Useful first-pass for procurement teams; final selection usually needs deeper eval
New to this calculator? Start with the ⚡ Playground — a few sliders, instant ballpark. Then switch to the 🧮 Calculator for your exact number.
Two ways to use this: visualize in the Playground, then get your number in the Calculator.
Which model tier should you use?
Quality and complexity choose the tier; volume scales it. One estimate from all three.
Recommended tier
—
Change the input sliders below to see the recommendation.
✓
Recommended tier—
💡Complexity sets the tier (1-3 budget, 4-6 mid-tier, 7+ premium). Quality is a floor: 7+ rules out budget, 9+ forces premium. Monthly volume multiplies the tier rate.
Tell us about your workload and we'll surface the cheapest model that can actually handle it — across every provider, at current prices.
📊 Not sure of a value? Fields with a ▾ Typical pill offer broad industry ballparks (sourced typical ranges) so you can move forward now — your result gets more accurate as you replace them with your own measured numbers. Values marked * are rough estimates.
📥 Inputs you provide
- Your roleTunes filter defaults to your priorities
- Your workload scenarioPre-tunes capability filters
- Capability chipsAND-filters by feature
- Max input price per 1M tokensCuts everything above your ceiling
- Minimum context windowCuts small-window models
- Provider filtersVendor allowlist
📤 Outputs you get
- Ranked model listModels matching all filters
- Side-by-side compareCheck 2-4 to compare in detail
- Pricing freshnessDays since last verification
- Capability tagsWhat this model can do
🎯 Use your results to
🎯
Shortlist in 60 seconds
Cut 150+ models to 5-10 candidates without reading any vendor blog post
🔍
Find what you didn't know existed
New providers launch monthly — this surfaces them instead of you finding out on Twitter
🛒
Make procurement defensible
Compare table is screenshot-ready for finance and security review
📦
Hand off to other calcs
Export shortlist directly to Cost Calculator or Multi-Model Router for the actual decision
1
Pick role + workload
Select your role (developer / PM / finance / etc.) and workload scenario. Filters auto-tune for what you actually need.
2
Apply capability chips
Toggle Vision / Audio / Video / Tier-1 / Flagship / Budget. Capability chips work as AND filters.
3
Dial in price + context
Slide max price (cuts pricy frontier) and min context (cuts small-window models). Watch the list collapse to 5-10 candidates.
4
Compare and ship
Checkbox 2-4 models, then export to Cost Calculator or Multi-Model Router for head-to-head at your workload.
1
Pick your role + workload first
The role dropdown adjusts default filter weights (developer prioritizes context + reasoning; finance prioritizes price). Workload scenario (chat, RAG, agent, code, content) pre-tunes the capability filters. These two choices alone often cut the list from 150+ to ~30 before you touch anything else.
2
Apply capability AND filters
Vision / Audio / Video filter to multimodal models. Tier-1 provider means OpenAI / Anthropic / Google only (rules out residency-sensitive vendors for regulated workloads). Flagship only hides budget variants. Budget chip caps at $1/M input. Provider chips give vendor-specific filtering. All chips compose with AND logic — toggling more narrows further.
3
Dial price + context windows
Max input price slider goes 0-$30/M — most production workloads should sit at $3-$10. Min context slider goes 0-2M tokens — set to your largest realistic prompt (1.2× your worst-case). The freshness column shows when each price was last verified; anything >30 days warrants a vendor-page double-check before signing a contract.
4
Compare 2-4, then export
Check 2-4 candidates and hit Compare to see them side-by-side. Then hand the shortlist to Cost Calculator for exact $/month at your token shape, or Multi-Model Router if you have mixed-complexity traffic (a cheaper model for easy queries, a smarter one for hard ones). Treat this as discovery only — actual cost decisions need a workload-specific run downstream.
👇 Now try the calculator below with your own AI workloads
👤 Your rolePre-tunes filter weights based on what your role typically cares about (developer = context + reasoning; PM = balance; finance = price + provider).How to choose: Pick the role closest to the decision you're making. If your work spans multiple roles, choose the one most cost-sensitive — it forces the tightest filter.Read the full guide →
Hint: Pick what you do; we tune the recommendations to it.
📊 Your workloadTells the finder what capabilities matter for your workload (RAG needs long-context; agents need reasoning + tool use; classification needs cheap + fast).How to choose: Pick the closest match. If you're building multiple things, run the finder once per workload — they often pick different optimal models.Read the full guide →
Hint: Pick your typical job; we filter to models that suit it.
🔀 Sort by
Providers:Per-provider toggle to include/exclude a vendor entirely. Useful when contract or compliance constraints limit which vendors you can use.How to choose: Start with everyone checked. Uncheck only when you have a concrete reason (no GDPR-compliant region, no enterprise BAA, existing volume commitment elsewhere).Read the full guide →
Hint: Limit to vendors you will actually use. Leave all on to compare widely.
Modality:Toggle chips to keep only models that support a feature. Vision filters multimodal vision-capable models; Tier-1 keeps OpenAI/Anthropic/Google only; Flagship hides budget variants.How to choose: Enable only chips for capabilities you actually need NOW. Adding "Vision" because you might want it eventually pollutes results with multimodal premium pricing.Read the full guide →
Hint: Switch on only what you need it to handle (images, audio).
Tags:
Compare:
Check boxes in the table to add up to 3 models for side-by-side comparison.
| Model | Provider | Input $/M ↓ | Output $/M | Cached $/M | Batch | Context | Modalities | Tags |
|---|
📋 What now?
- Shortlist your exact-fit models — filter by role, workload, price, and context
- Compare 2–3 finalists — side-by-side price, context, and capabilities
- Price your real workload — hit "Calc →" on any row to drop it into the cost calculator
Need help choosing or cutting your AI bill?
💼 Talk to a CloudIntelligence advisor →
Now that you have your shortlist…
What this means + what to do next
💡 What to consider beyond this shortlist for full TCO
- Real workload cost — sticker price tells you nothing about your token shape
- Quality at YOUR task — same model is great on one workload, weak on another
- Cache-aware effective pricing — caching changes the effective $/M dramatically for repeat-context workloads
- Vendor lock-in cost — switching prompts between vendors often takes weeks of eval
Rule of thumb: Treat this as discovery only. Real selection requires running 2-3 shortlisted models against an eval set on YOUR task, then costing them at YOUR token shape.
Quantify the hidden costs:
- Get exact $/month at your token shape per candidate model Cost Calculator
- Quantifies lock-in risk before you commit Vendor Concentration Risk
- Some models support prompt caching — drops effective $/M by 50-80% above ~22% hit rate Prompt Cache Roi
$ How this fits your overall ROI
This is a discovery tool. ROI conversations happen downstream:
- Is the cheapest model in the shortlist quality-acceptable for my task? (Run eval.)
- Is the most-capable model in my shortlist 2× better, or 10× better? (The price gap usually reflects 2×, the value gap often reflects 1.3×.)
- How locked in am I to my current vendor — what would migration cost?
Bridge to ROI:
- Mixed workloads benefit from routing — different models for different query types Multi Model Router
- For some open-weight models, self-hosting becomes cheaper above a usage threshold Self Host Breakeven
- Translate $/request into $/customer or $/feature margin Margin Calculator
Doing something different?
If you already have specific candidates in mind, skip discovery:
- You know which model you want; you just need the dollar number Cost Calculator
- You want the cheapest model meeting a quality floor for a specific workload Cheapest Model
- You've decided you want multiple models for different traffic patterns Multi Model Router
Need help using this calculator for your workloads?
AICost.ai has 50+ calculators and playbooks. Schedule an AvatarVA meeting and we'll work through your real cost scenarios across AI & Cloud: visibility, cost reduction, optimization, forecasting and capacity planning, without sacrificing accuracy or performance.
📅 Schedule an AvatarVA meeting →📖 Data sources & methodology
Methodology
- All prices are USD per 1 million tokens, current as of 2026-07-28.
- Vendor-published values have no mark. Inferred/extrapolated values are marked with * and listed below.
- Batch API discounts are 50% off standard rates across providers that offer Batch mode.
- Prompt caching discounts vary by provider (typically 80-90% off cached input tokens).
- Regional data-residency surcharges (Anthropic 1.1x, OpenAI 1.1x, Google regional tiers) are NOT included in base rates.
- Long-context pricing tiers apply when input exceeds model threshold.
- Embedding prices are input-only (no output tokens generated).
Primary sources
Last-verified date is the most recent successful daily snapshot
(aicost_pricing_snapshots) or, when no snapshot exists yet,
the latest successful crawler run (aicost_crawler_runs).
10 of 10
vendors are currently verified. Aggregator services (TokenCost, AI Pricing Guru, etc.)
are not listed.
Anthropic
2026-07-28
Daily snapshot since Sep 2023 · 631 days captured
Anthropic Docs
2026-07-28
Daily snapshot since Sep 2023 · 631 days captured
Google AI
2026-07-28
Daily snapshot since Dec 2023 · 607 days captured
Google Vertex
2026-07-28
Daily snapshot since Dec 2023 · 607 days captured
DeepSeek
2026-07-28
Daily snapshot since May 2024 · 546 days captured
Voyage AI
2026-07-28
Inferred values (marked with * in calculator tables)
Derived from industry conventions, not directly published by the vendor. Typical conventions: cached input = 10% of base (90% off), Batch API = 50% of base (50% off).
| Vendor / Model | Field | Why it’s inferred |
|---|---|---|
| Anthropic — Claude Sonnet 4.6 | cachedInput |
Derived at 10% of input rate — Anthropic publishes 90% cache-hit discount on this tier. |
| Anthropic — Claude Sonnet 4.5 | cachedInput |
Derived at 10% of input rate; same 90% cache-hit convention as Sonnet 4.6. |
| Anthropic — Claude Sonnet 4.5 | batchInput |
Derived at 50% of standard input — Anthropic documents uniform 50% Batch discount. |
| Anthropic — Claude Sonnet 4.5 | batchOutput |
Derived at 50% of standard output — Anthropic documents uniform 50% Batch discount. |
| Anthropic — Claude Haiku 4.5 | cachedInput |
Derived at 10% of input rate — Anthropic 90% cache-hit discount convention. |
| OpenAI — GPT-5.4 Mini | cachedInput |
Derived at 10% of input — OpenAI documents automatic 90% discount on cache hits across GPT-5.x tier. |
| OpenAI — GPT-5.4 Nano | cachedInput |
Derived at 10% of input — OpenAI 90% cache-hit convention. |
| OpenAI — GPT-5.4 Nano | batchInput |
Derived at 50% of input — OpenAI Batch API uniform 50% discount. |
| OpenAI — GPT-5.4 Nano | batchOutput |
Derived at 50% of output — OpenAI Batch API uniform 50% discount. |
| OpenAI — GPT-5.4 Pro | cachedInput |
Derived at 10% of input — OpenAI 90% cache-hit convention. |
| OpenAI — GPT-5.4 Pro | batchInput |
Derived at 50% of input — OpenAI Batch API uniform 50% discount. |
| OpenAI — GPT-5.4 Pro | batchOutput |
Derived at 50% of output — OpenAI Batch API uniform 50% discount. |
| OpenAI — GPT-5.2 | cachedInput |
Derived at 10% of input; no residency uplift. |
| OpenAI — GPT-5.2 | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5.2 | batchOutput |
Derived at 50% of output. |
| OpenAI — GPT-5 | cachedInput |
Derived at 10% of input. |
| OpenAI — GPT-5 | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5 | batchOutput |
Derived at 50% of output. |
| OpenAI — GPT-5.5 Pro | cachedInput |
Derived at 10% of input — OpenAI does not publish a cached rate for *-pro models; using the family convention. |
| OpenAI — GPT-5.5 Pro | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5.5 Pro | batchOutput |
Derived at 50% of output. |
| OpenAI — GPT-5.2 Pro | cachedInput |
Derived at 10% of input — pro-tier convention. |
| OpenAI — GPT-5.2 Pro | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5.2 Pro | batchOutput |
Derived at 50% of output. |
| OpenAI — GPT-5.1 | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5.1 | batchOutput |
Derived at 50% of output. |
| OpenAI — GPT-5 Pro | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5 Pro | batchOutput |
Derived at 50% of output. |
| OpenAI — GPT-5 Nano | cachedInput |
Derived at 10% of input. |
| OpenAI — GPT-5 Nano | batchInput |
Derived at 50% of input. |
| OpenAI — GPT-5 Nano | batchOutput |
Derived at 50% of output. |
| Google — Gemini 3 Flash | cachedInput |
Derived at 10% of input — Google caching discount convention ~90%. |
| Google — Gemini 3.1 Flash-Lite | cachedInput |
Derived at 10% of input — Google caching convention. |
| Google — Gemini 3.1 Flash-Lite | batchInput |
Derived at 50% of input — Google Batch API uniform 50% discount. |
| Google — Gemini 3.1 Flash-Lite | batchOutput |
Derived at 50% of output — Google Batch API uniform 50% discount. |
| Google — Gemini 2.5 Pro | cachedInput |
Derived at 10% of input. |
| Google — Gemini 2.5 Flash | cachedInput |
Derived at 10% of input. |
| Google — Gemini 2.5 Flash-Lite | cachedInput |
Derived at 10% of input — Google caching convention. |
| Google — Gemini 2.5 Flash-Lite | batchInput |
Derived at 50% of input — Google Batch API uniform 50% discount. |
| Google — Gemini 2.5 Flash-Lite | batchOutput |
Derived at 50% of output — Google Batch API uniform 50% discount. |
| Google — Gemini 2.0 Flash | cachedInput |
Derived at 25% of input per Google 2.0 family caching rates. |
| Google — Gemini 2.0 Flash | batchInput |
Derived at 50% of input — Google Batch API uniform 50% discount. |
| Google — Gemini 2.0 Flash | batchOutput |
Derived at 50% of output — Google Batch API uniform 50% discount. |
| Google — Gemini 2.0 Flash-Lite | cachedInput |
Derived at 10% of input — Google caching convention. |
| Google — Gemini 2.0 Flash-Lite | batchInput |
Derived at 50% of input — Google Batch API uniform 50% discount. |
| Google — Gemini 2.0 Flash-Lite | batchOutput |
Derived at 50% of output — Google Batch API uniform 50% discount. |
| xAI — Grok 4 (legacy) | cachedInput |
Extrapolated at 25% of base. |
Pricing is cross-verified against the
LiteLLM community registry
when available. Daily snapshots are kept in aicost_pricing_snapshots;
every change is logged to aicost_price_changelog with old & new
values for full audit trail. Read the full methodology →