Pricing verified July 30, 2026/10 current models
One dashboard.
Every model bill.
Track OpenAI, Claude, Gemini, and custom clients live—then compare cost per successful task, not merely cost per token.
01
Shape your workload
Every figure is per request unless noted.
ProcessingBatch is async and 50% off eligible token rates.
NoteRequests above 272K input tokens use 2× input and 1.5× output rates for the full request.
02
Same workload. Every model.
Sorted by your estimated monthly bill—not a marketing benchmark.
12K calls · 18K in · 2K out
| Model | Input / MTok | Cache read | Output / MTok | Your monthly cost |
|---|---|---|---|---|
G Gemini 3.5 Flash-Litegemini-3.5-flash-lite Cheapest | $0.30 | $0.030 | $2.50 | $98.88 |
A Claude Haiku 4.5claude-haiku-4-5 | $1.00 | $0.100 | $5.00 | $250 |
O GPT-5.6 Lunagpt-5.6-luna | $1.00 | $0.100 | $6.00 | $274 |
G Gemini 3.5 Flashgemini-3.5-flash | $1.50 | $0.150 | $9.00 | $410 |
A Claude Sonnet 5claude-sonnet-5 | $2.00 | $0.200 | $10.00 | $499 |
G Gemini 3.1 Pro Previewgemini-3.1-pro-preview | $2.00 | $0.200 | $12.00 | $547 |
O GPT-5.6 Terragpt-5.6-terra | $2.50 | $0.250 | $15.00 | $684 |
A Claude Opus 4.8claude-opus-4-8 | $5.00 | $0.500 | $25.00 | $1,248 |
O GPT-5.6 Solgpt-5.6-sol | $5.00 | $0.500 | $30.00 | $1,368 |
A Claude Fable 5claude-fable-5 | $10.00 | $1.000 | $50.00 | $2,496 |
Click a row to load it in the calculator. Rates exclude tool calls, storage, fine-tuning, and provider-specific taxes.
03
Your usage. Better options.
Combine providers, attribute spend by project, and see the lowest list-price model that fits the same token workload.
Usage importRuns in your browser
Drop in one clean CSV.
One row per project and model. Token totals can come from any provider export; an optional reported cost becomes the baseline.
Required
model · requests · input_tokens · output_tokens◎
Private by design. The file is parsed locally and never uploaded to TokenTab.
Demo workspace3 valid rows
Cross-provider snapshotReported + modeled USD
Spend by providerOne view / all vendors
OOpenAI$4,712
AAnthropic$1,086
GGoogle$1,166
| Project | Current model | Requests | Current spend | Lowest compatible price | Potential saving |
|---|---|---|---|---|---|
Support copilot1 model | GPT-5.6 Terra | 82,000 | $4,712 | Gemini 3.5 Flash-Lite$676 | Save $4,037 |
Batch enrichment1 model | Gemini 3.5 Flash | 140,000 | $1,166 | Gemini 3.5 Flash-Lite$272 | Save $893 |
Docs search1 model | Claude Sonnet 5 | 28,000 | $1,086 | Gemini 3.5 Flash-Lite$219 | Save $867 |
Price comparison preserves each row’s request volume, cache usage, batch mode, and context requirements. Launch the local dashboard to add private outcome telemetry and verify quality before switching.
04One command.
One command.
Full dashboard.
The zero-account CLI wraps JavaScript and Python clients, syncs separately authorized providers, runs consent-gated evals, and opens this interface at 127.0.0.1. Keys and model content never enter our site.
Loopback onlyKeys never savedMIT licensed
$ npx --yes --package=github:adididitagain/tokentab#v1.0.0 tokentab dashboard
ENGINElocal · privateLIVE SDKresponse usageEFFICIENCYcost / success
$ npm install https://github.com/adididitagain/tokentab/releases/download/v1.0.0/tokentab-ai-1.0.0.tgz openai
$ tokentab eval --file evals.json --models gpt-5.6-terra,claude-sonnet-5 --allow-provider-data
05
Pricing you can audit.
No mystery multiplier. The formula and sources are visible, and every rate is stored as USD per million tokens.
((input − cached) × input rate)
+ (cached × cache rate)
+ (output × output rate)- Cache-read token rates
- 50% batch processing
- Long-context thresholds
- Shareable scenario URLs
- Local usage optimization