Calculator
Kalkulator Biaya Fine-tuning LLM
Hitung biaya fine-tuning — token training × tarif per juta, plus uplift per token untuk inference model custom.
Pricing refreshed: 2026-05-01
Total tahun 1 · cheapest
Fireworks · Llama 4 8B
$248
| Provider | Base model | Biaya training | Inference bulanan | Total tahun 1 |
|---|---|---|---|---|
| Fireworks | Llama 4 8B | $8 | $20 | $248 |
| Together | Llama 4 8B LoRA adapter; full fine-tune more | $15 | $22 | $279 |
| Cohere | Command R | $30 | $48 | $606 |
| OpenAI | GPT-4o mini Inference is 2× base mini rate | $45 | $48 | $621 |
| Mistral | Mistral Small 3 $2/mo hosting per deployed adapter | $45 | $58 | $741 |
| Fireworks | Llama 4 70B | $45 | $90 | $1,125 |
| Together | Llama 3.3 70B | $75 | $88 | $1,131 |
| OpenAI | GPT-5 mini | $60 | $96 | $1,212 |
| Together | Llama 4 70B | $90 | $120 | $1,530 |
| OpenAI | o3-mini | $75 | $136 | $1,707 |
| AWS Bedrock | Claude Haiku 4.5 (custom) Provisioned throughput required | $120 | $303 | $3,756 |
| Mistral | Mistral Large 2 | $135 | $564 | $6,903 |
| OpenAI | GPT-4o Inference is 1.5× base GPT-4o rate | $375 | $600 | $7,575 |
Training cost = tokens × epochs × per-million rate. Inference uses the fine-tuned model's uplifted per-token rate, which is always higher than the base model. Year-1 total = one-time training + 12 months of inference.
Pertanyaan yang sering diajukan
Seberapa akurat kalkulator ini?+
Harga bersumber dari dokumentasi resmi provider dan diperbarui bulanan. Tagihan nyata bisa berbeda 5–15%.
Apakah harga dalam USD?+
Ya, semua harga dalam USD sesuai mata uang penagihan provider.
Seberapa sering data diperbarui?+
Tabel harga ditinjau dan diperbarui setiap tanggal 1.
Bisakah saya andalkan ini untuk anggaran?+
Gunakan sebagai estimasi. Untuk anggaran produksi, validasi dengan pilot 1 minggu.