Llama 3.1 8B Instruct API pricing
Meta · Budget price tier · Llama 3.x · released 2024-07. Prices in USD per 1 million tokens. Data snapshot: 2026-11-21.
A support ticket workload of 100K requests a month costs $5.00 on Llama 3.1 8B Instruct; the next cheaper model in the Budget price tier for this workload, Ling 3.0 Tiny, does it for $4.50. The price we quote for Llama 3.1 8B Instruct has changed 3 times since 2026-08-19, most recently on 2026-11-09 when the quoted host changed (-52.1%); it is now the 4th cheapest of 160 models in the Budget price tier. None of its 4 hosts undercuts the price we quote. It is rated a stable release but does not meet all four criteria of the FinOps Friendly badge.
- Input: $0.02 per 1M tokens
- Output: $0.04 per 1M tokens
- Batch API (in/out): Not published
- Prompt-cache read: Not published
- Context window: 131K tokens
- Arena ELO: 1187 as of 2026-11-11
Available on
AWS Bedrock regions: us-east-1, us-east-2, us-west-2
Market position
On blended price ($0.034 per 1M tokens at 30% input, 70% output) it is cheaper than 98% of models in the Budget price tier: 4th cheapest of 160. Across every price tier it is cheaper than 99% of priced text models in the catalogue: 4th cheapest of 308. Its Arena ELO of 1187 ranks 154th of the 157 models that carry one, tied with one other, above 1% of them. With an Arena score of 1187, Llama 3.1 8B Instruct sits below OptimNow's lowest working quality threshold, Standard (Arena 1419+), so the quality-bar method would not shortlist it for any of the 8 use cases the table prices. Each threshold is a distance below the best ranked model, which scores 1529 on Arena scores as of 11 Nov 2026.
Licence, size and capabilities
Llama 3.1 8B 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 8B parameters. Listed capabilities: Text, Agents. Filling its 131K context window once costs $0.00262 in input tokens at list price.
Pricing by host
A host is a company that runs the model and sells it by the token: the model maker, a cloud platform like Amazon Bedrock, Azure or Vertex, or an independent host. One model, several hosts, and often several prices. This model is sold by 4 different hosts. The price we quote elsewhere on this page is marked. We quote one host's rate so that models stay comparable; it is not a recommendation to buy there.
The price we quote is already the cheapest of the 4 hosts: DeepInfra at $0.02 input / $0.04 output per 1M tokens. It serves a quantized fp8 build, which is not the same product as the full-precision model. The dearest, CoreWeave, charges 6.5x the cheapest.
| Host | Input $/1M | Output $/1M | Precision |
|---|---|---|---|
| DeepInfra (price we quote) | $0.02 | $0.04 | fp8 |
| Groq | $0.05 | $0.08 | not declared |
| Amazon Bedrock | $0.22 | $0.22 | not declared |
| CoreWeave | $0.22 | $0.22 | bf16 |
List prices as published on 2026-11-21, 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.
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 Meta 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 | $5 |
| Knowledge Q&A | 2,000 / 800 | $7 |
| Meeting Summary | 10,000 / 1,200 | $25 |
| Marketing Content | 2,500 / 1,800 | $12 |
| Coding Task | 3,000 / 2,000 | $14 |
| Invoice Processing | 1,500 / 600 | $5 |
| Call Summary | 2,000 / 700 | $7 |
| Agent Workflow | 6,000 / 3,000 | $24 |
Price history
Llama 3.1 8B Instruct entered our daily price archive on 2026-08-19; tracking began on 2026-08-15. It listed at $0.05 input / $0.08 output per 1M tokens that day. The price we quote has changed 3 times since then, every time through a change of quoted host rather than a repricing. The latest change, on 2026-11-09, was -52.1% on the blended price (30% input, 70% output), to $0.02 input / $0.04 output (quoted host changed, not a repricing). Net of all changes, the blended price is -52.1% against 2026-08-19.
| Date | Old price (in / out) | New price (in / out) | Blended change |
|---|---|---|---|
| 2026-11-09 | $0.05 / $0.08 | $0.02 / $0.04 | -52.1% (quoted host changed, not a repricing) |
| 2026-11-01 | $0.02 / $0.04 | $0.05 / $0.08 | +108.8% (quoted host changed, not a repricing) |
| 2026-10-28 | $0.05 / $0.08 | $0.02 / $0.04 | -52.1% (quoted host changed, not a repricing) |
Alternatives to Llama 3.1 8B Instruct
None of the 4 comparable models below costs less than Llama 3.1 8B Instruct on the blended price.
- Next cheaper model in the Budget price tier: Mistral Nemo (Mistral) at $0.0297 blended, 13% less.
- Next more expensive: Ling 3.0 Tiny (inclusionai) at $0.036 blended, 6% more.
The nearest models in the Budget price tier on Arena ELO, each with its blended price and score difference against Llama 3.1 8B Instruct.
- Command R (08-2024) (Cohere): $0.15 in / $0.60 out per 1M, Arena ELO 1187 (1268% more expensive, same Arena ELO)
- Nova Micro 1.0 (Amazon): $0.035 in / $0.14 out per 1M, Arena ELO 1208 (219% more expensive, 21 ELO points higher)
- Phi 4 (Microsoft): $0.07 in / $0.14 out per 1M, Arena ELO 1217 (250% more expensive, 30 ELO points higher)
- Nova Lite 1.0 (Amazon): $0.06 in / $0.24 out per 1M, Arena ELO 1229 (447% more expensive, 42 ELO points higher)
- Cheapest within 15 Arena points: Llama 3.1 8B Instruct is itself the cheapest of the 2 models scoring within 15 points of its ELO.
Other models from Meta
The 8 nearest to Llama 3.1 8B Instruct in price, of 15 other Meta models in the catalogue. Prices per 1M tokens, input then output.
- Muse Spark 1.2 Contributor: $0.10 / $0.20
- Muse Spark 1.3 Contributor: $0.10 / $0.20
- Llama Guard 4 12B: $0.18 / $0.18
- Llama 4 Scout: $0.10 / $0.30
- Llama 3.2 3B Instruct: $0.05 / $0.33
- Muse Ember: $0.10 / $0.50
- Llama 3.3 70B Instruct: $0.71 / $0.71
- Llama 3.1 70B Instruct: $0.72 / $0.72
Common questions about Llama 3.1 8B Instruct pricing
- How much does Llama 3.1 8B Instruct cost?
- Llama 3.1 8B Instruct costs $0.02 per 1M input tokens and $0.04 per 1M output tokens at DeepInfra, the host whose rate we quote. DeepInfra serves a quantized fp8 build, which is not the same product as the full-precision model. Prices as listed on 2026-11-21.
- Which host is cheapest for Llama 3.1 8B Instruct?
- The price we quote is already the cheapest of the 4 hosts: DeepInfra at $0.02 input / $0.04 output per 1M tokens. It serves a quantized fp8 build, which is not the same product as the full-precision model. The dearest, CoreWeave, charges 6.5x the cheapest. Host rates as listed on 2026-11-21.
- Does Llama 3.1 8B Instruct offer batch or cached pricing?
- No. Neither a batch rate nor a prompt-cache rate is published for Llama 3.1 8B Instruct as of 2026-11-21.
- What would Llama 3.1 8B Instruct cost per month?
- With the Support Ticket profile (1,500 input and 500 output tokens per request), 100,000 requests a month cost about $5 on Llama 3.1 8B Instruct at the list price quoted on 2026-11-21.
- Can Llama 3.1 8B Instruct be self-hosted?
- Yes, within the terms of its licence. Llama 3.1 8B 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.
- Has the price of Llama 3.1 8B Instruct changed?
- Llama 3.1 8B Instruct entered our daily price archive on 2026-08-19; tracking began on 2026-08-15. It listed at $0.05 input / $0.08 output per 1M tokens that day. The price we quote has changed 3 times since then, every time through a change of quoted host rather than a repricing. The latest change, on 2026-11-09, was -52.1% on the blended price (30% input, 70% output), to $0.02 input / $0.04 output (quoted host changed, not a repricing). Net of all changes, the blended price is -52.1% against 2026-08-19.
Its weights are published, which changes what you may run yourself: the cost gap between open weights and proprietary APIs.
These are list prices. Your real cost also depends on caching and batching, negotiated discounts, marketplace agreements, and the contracts you already have. OptimNow audits exactly that.
Build the business case · Get a FinOps review
LLM pricing · Models by maker · Cloud compute pricing · Compare models · Price barometer · FinOps guides · API and MCP server