FM
FrontierModel.app
Frontier-model developer tools · real tokenizer · live cost data
Tokenizer & cost lab
Paste a prompt to see exact token boundaries, then compare what it costs to run across every model in the survey — using each lab’s own published pricing.
Prompt
Chars 0
Words 0
Tokens 0
Tokens/word 0
Token boundaries
Type or paste a prompt to see real token boundaries and live cost across every model.
Loading tokenizer…
Cost to run this prompt across every model
| Model | Tier | In $/1M | Out $/1M | Input cost | Output cost | Total × runs |
|---|
Input/output cost is per single run at each model’s published price; the last column multiplies the per-run total by your run count. n/p = price not published (shown as such rather than as free).
Model comparison matrix
Every model in the Survey of the Frontier, sortable and filterable. Click a model to read its full assessment.
| # | Model | Lab | Tier | Context | In $/1M | Out $/1M | AA index | License |
|---|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 5 | Anthropic | Flagship | ~1M | $5.00 | $25.00 | 61 independent | closed |
| 2 | Claude Fable 5 | Anthropic | Flagship | ~1M | $10.00 | n/a | 60 independent | closed |
| 3 | GPT-5.6 Sol | OpenAI | Flagship | 1.05M | $5.00 | $30.00 | 59 independent | closed |
| 4 | Gemini 3.1 Pro | Flagship | ~1M | $2.00 | $12.00 | not surveyed | closed | |
| 5 | Grok 4.6 | xAI | Flagship | ~500K | n/a | n/a | not surveyed | closed |
| 6 | GPT-5.6 Terra | OpenAI | Balanced | 1.05M | $2.00 | $12.00 | not surveyed | closed |
| 7 | Claude Sonnet 5 | Anthropic | Balanced | ~1M | $2.00 | n/a | not surveyed | closed |
| 8 | Gemini 3.5 Flash | Balanced | 1M | $1.50 | $9.00 | not surveyed | closed | |
| 9 | Grok 4.5 | xAI | Balanced | ~500K | $2.00 | $6.00 | 54 independent | closed |
| 10 | GPT-5.6 Luna | OpenAI | Efficient | 1.05M | $0.20 | $1.20 | not surveyed | closed |
| 11 | Claude Haiku 4.5 | Anthropic | Efficient | ~200K | n/a | n/a | not surveyed | closed |
| 12 | DeepSeek V4-Pro | DeepSeek | Open | 1M | $0.44 | $0.87 | 53 independent | MIT |
| 13 | DeepSeek V4-Flash | DeepSeek | Open | 1M | $0.14 | $0.28 | not surveyed | MIT |
| 14 | Kimi K3 | Moonshot AI | Open | ~1M | $0.60 | $3.00 | not surveyed | Modified MIT |
| 15 | Qwen3.8-Max | Alibaba | Open | 1M | $2.00 | $6.00 | not surveyed | open |
| 16 | GLM-5.2 | Z.ai | Open | 1M | n/a | n/a | not surveyed | open |
AA index = Artificial Analysis intelligence score. independent = independently measured; vendor = vendor-reported or not yet independently surveyed.
Data live from frontiermodel.io · the Survey of the Frontier.
Tokenizer:
o200k_base (GPT-4o/5 class) · BYOK live arena coming later.