Llama 3.3 70B Instruct API Pricing
Meta · Budget · Llama 3.x · released 2024-12. Prices in USD per 1 million tokens. Data snapshot: 2026-10-03.
The price we publish for Llama 3.3 70B Instruct has changed 2 times since 2026-08-19, most recently on 2026-09-03 when the reference host changed (-64.2%); it is now the 52nd cheapest of 145 Budget models. None of its 10 hosts undercuts the price quoted here. It is rated a stable release but does not meet all four criteria of the FinOps Friendly badge.
- Input: $0.10 per 1M tokens
- Output: $0.32 per 1M tokens
- Batch API (input/output): not published
- Prompt-cache read: not published
- Context window: 131K
- Arena ELO: 1318 (as of 2026-09-25)
Llama 3.3 70B Instruct market position
On blended price ($0.254 per 1M tokens at 30% input, 70% output) it is cheaper than 65% of Budget models: 52nd cheapest of 145. Across every category it is cheaper than 82% of priced text models in the catalogue: 52nd cheapest of 285. Its Arena ELO of 1318 ranks 63rd of the 68 models that carry one, tied with one other, above 6% of them.
Llama 3.3 70B Instruct licence, size and capabilities
Llama 3.3 70B Instruct is open weights, which is not the same as open source: the weights are published, but under Llama 3.x, a licence that restricts some uses. Read it before self-hosting commercially. Its published size is 70B parameters. Listed capabilities: Text, Agents. Filling its 131K context window once costs $0.01 in input tokens at list price.
Monthly cost at 100K requests, by business use case
From the after-discounts figure (prompt caching on the cached input share, batch API for async workloads, only where the provider publishes those rates), which assumes you implement both, to the list price. One figure means no discount is published.
| Use case | Tokens (in/out) | Monthly (after discounts – list) |
|---|---|---|
| Support Ticket | 1,500 / 500 | $31 |
| Knowledge Q&A | 2,000 / 800 | $46 |
| Meeting Summary | 10,000 / 1,200 | $138 |
| Marketing Content | 2,500 / 1,800 | $83 |
| Coding Task | 3,000 / 2,000 | $94 |
| Invoice Processing | 1,500 / 600 | $34 |
| Call Summary | 2,000 / 700 | $42 |
| Agent Workflow | 6,000 / 3,000 | $156 |
Llama 3.3 70B Instruct pricing by host
A host is a company that runs the model and sells it by the token — the lab that made it, a cloud platform like Amazon Bedrock, Azure or Vertex, or an independent provider. One model, several sellers, and often several prices. This model is sold by 10 different hosts. The price quoted above is the one we publish so that models stay comparable; it is not necessarily the cheapest way to buy this model. List prices as published on 2026-10-03 via OpenRouter — negotiated rates, committed-use discounts and provisioned capacity are not shown. These hosts do not all serve the model at the same numeric precision, so the cheapest is not a like-for-like substitute for the dearest.
The price we publish is already the cheapest of the 10 hosts: DeepInfra at $0.10 input / $0.32 output per 1M tokens. It serves a quantized fp8 build, which is not the same product as the full-precision model. The dearest, Cloudflare, charges 6.6x the cheapest.
| Host | Input $/1M | Output $/1M |
|---|---|---|
| DeepInfra (the price shown above) — fp8 | $0.10 | $0.32 |
| Novita — bf16 | $0.14 | $0.40 |
| Parasail — fp8 | $0.22 | $0.50 |
| AkashML — fp8 | $0.20 | $0.52 |
| CoreWeave — fp16 | $0.71 | $0.71 |
| $0.72 | $0.72 | |
| Groq | $0.59 | $0.79 |
| SambaNova | $0.45 | $0.90 |
| Together | $1.04 | $1.04 |
| Cloudflare — fp8 | $0.29 | $2.25 |
Llama 3.3 70B Instruct price history
Llama 3.3 70B Instruct entered our daily price archive on 2026-08-19, 2026-08-15 being the first day we tracked any model. It listed at $0.10 input / $0.32 output per 1M tokens that day. The price we publish has changed 2 times since then, 1 of them a change of reference host rather than a repricing. The latest change, on 2026-09-03, was -64.2% on the blended price (30% input, 70% output), to $0.10 input / $0.32 output (reference host changed, not a repricing). Net of all changes, the blended price is back where it started on 2026-08-19.
| Date | Old price (in / out) | New price (in / out) | Blended change |
|---|---|---|---|
| 2026-09-03 | $0.71 / $0.71 | $0.10 / $0.32 | -64.2% (reference host changed, not a repricing) |
| 2026-08-26 | $0.10 / $0.32 | $0.71 / $0.71 | +179.5% |
Alternatives to Llama 3.3 70B Instruct
- Next cheaper Budget model: Llama 4 Scout (Meta) at $0.240 blended, 6% less.
- Next more expensive: Hy3 (Tencent) at $0.256 blended, 1% more.
The nearest Budget models on Arena ELO, each with its blended price and score difference against Llama 3.3 70B Instruct.
- GPT-4o-mini (OpenAI) — $0.15 in / $0.60 out per 1M, Arena ELO 1318: 83% more expensive, same Arena ELO
- Llama 4 Scout (Meta) — $0.10 in / $0.30 out per 1M, Arena ELO 1322: 6% cheaper, 4 ELO points higher
- Llama 4 Maverick (Meta) — $0.19 in / $0.65 out per 1M, Arena ELO 1327: 102% more expensive, 9 ELO points higher
- Llama 3.1 70B Instruct (Meta) — $0.40 in / $0.40 out per 1M, Arena ELO 1293: 57% more expensive, 25 ELO points lower
- Cheapest within 15 Arena points: Llama 4 Scout (Meta), 6% cheaper, 4 ELO points higher, out of 5 models in that band.
- Highest Arena ELO within 20% of this price: Llama 4 Scout (Meta), 6% cheaper, 4 ELO points higher, out of 2 models in that price band.
Common questions about Llama 3.3 70B Instruct pricing
- How much does Llama 3.3 70B Instruct cost?
- Llama 3.3 70B Instruct costs $0.10 per million input tokens and $0.32 per million output tokens on the API.
- Does Llama 3.3 70B Instruct offer batch or cached pricing?
- No batch or cache pricing is published for this model.
- What would Llama 3.3 70B Instruct cost per month?
- For a support-ticket workload (1,500 input + 500 output tokens per request) at 100K requests/month, Llama 3.3 70B Instruct costs about $31 at list prices.
- Has the price of Llama 3.3 70B Instruct changed?
- Llama 3.3 70B Instruct entered our daily price archive on 2026-08-19, 2026-08-15 being the first day we tracked any model. It listed at $0.10 input / $0.32 output per 1M tokens that day. The price we publish has changed 2 times since then, 1 of them a change of reference host rather than a repricing. The latest change, on 2026-09-03, was -64.2% on the blended price (30% input, 70% output), to $0.10 input / $0.32 output (reference host changed, not a repricing). Net of all changes, the blended price is back where it started on 2026-08-19.
These are list prices: real cost also depends on caching, batching, negotiated discounts, marketplace agreements, and existing contracts. Get a FinOps review from OptimNow or compare all models.
Build the business case for Llama 3.3 70B Instruct in the AI ROI Calculator: add your business value on top of these costs to get ROI, break-even and payback.
All models · Compare side by side · Cloud compute pricing · Price barometer · FinOps guides · API and MCP server