Llama 3.3 70B Instruct pricing calculator
Meta's popular dense 70B open-weight model, a common self-hosting baseline. Pricing shown is DeepInfra's serverless rate.
- Input
- $0.10 / 1M
- Cached input
- —
- Output
- $0.32 / 1M
- Context window
- 128K
- Batch API
- No
- Reasoning
- No
Workload
Quick presets
1,000 tokens ≈ 750 English words.
Pro / Agent mode
Caching, batch, multi-turn loops, reasoning tokens
Cost comparison
Cheapest / month
$5.86
Mistral Small 3.2 24B
Most expensive / month
$8.46
Llama 3.3 70B Instruct
Tokens / request
1.5K in
400 out (cheapest model)
| Remove | |||||
|---|---|---|---|---|---|
| Mistral Small 3.2 24BCheapestMistral via DeepInfra · $0.075 / $0.20 | $0.00019 | $0.1925 | $5.86 | $70.26 | |
| DeepSeek V4 Flash (open weights)DeepSeek via DeepInfra · $0.09 / $0.18 | $0.00021 | $0.207 | $6.30 | $75.55 | |
| Qwen3 32BAlibaba via DeepInfra · $0.08 / $0.28 | $0.00023 | $0.232 | $7.06 | $84.68 | |
| Llama 3.3 70B InstructMeta via DeepInfra · $0.10 / $0.32 | $0.00028 | $0.278 | $8.46 | $101.47 |
Monthly = daily × 30.42. Prices are list prices for the standard context tier. Long-context surcharges, regional pricing and taxes are not included.Official pricing ↗
What Llama 3.3 70B Instruct costs for common workloads
Monthly estimates using the presets above, at list price.
| Workload | Per request / session | Monthly | Yearly |
|---|---|---|---|
| Small Chatbot1K chats/day, short prompts | $0.00016 | $4.75 | $56.94 |
| RAG App5K queries/day, 4K-token retrieved context | $0.00053 | $80.31 | $963.60 |
| Coding Agent200 sessions/day × 25 turns | $0.1278 | $777.54 | $9,329 |
| Enterprise Scale250K requests/day, batched | $0.00036 | $2,738 | $32,850 |
Official pricing ↗ (Meta via DeepInfra) · API model ID meta-llama/Llama-3.3-70B-Instruct-Turbo
Llama 3.3 70B Instruct pricing FAQ
How much does Llama 3.3 70B Instruct cost per 1M tokens?
Llama 3.3 70B Instruct via DeepInfra costs $0.10 per 1M input tokens and $0.32 per 1M output tokens. Prices were last verified on Nov 21, 2026.
How much does one Llama 3.3 70B Instruct request cost?
A typical request with 1,000 input tokens and 500 output tokens costs about $0.00026. At 1,000 such requests per day that is $7.91 per month.
Does Llama 3.3 70B Instruct support prompt caching?
Llama 3.3 70B Instruct via DeepInfra has no discounted cached-input price listed, so every input token is billed at the full rate.
Is there a Batch API discount for Llama 3.3 70B Instruct?
No 50% batch discount is listed for Llama 3.3 70B Instruct via DeepInfra.
What is the context window of Llama 3.3 70B Instruct?
Llama 3.3 70B Instruct supports up to 128K tokens of context.
Is Llama 3.3 70B Instruct cheaper than Qwen3 32B?
For a 1,000-in / 500-out request, Llama 3.3 70B Instruct costs $0.00026 and Qwen3 32B costs $0.00022, so Qwen3 32B is 15% cheaper. Quality, speed and caching behaviour differ, so compare on your own workload above.
More AI cost tools
LLM API & Agent Cost Calculator
Compare per-request, monthly and yearly cost across 30+ models, with caching, batch and multi-turn agent modelling.
Open calculator →GBGPU & VRAM Calculator
See exactly how much VRAM a model needs (weights + KV cache + overhead) and which GPUs or Macs it fits on.
Open calculator →$/minVoice Agent Cost Calculator
Cost per minute and per call for STT + LLM + TTS + telephony voice agent stacks.
Open calculator →break-evenAPI vs Self-Hosting Break-Even
Find the monthly token volume where renting a dedicated GPU beats paying per token.
Open calculator →