This model is deprecated. We only continue performance benchmarking for the default 10k input token workload. Results for other workloads are historical and no longer updated.

Meta has launched a newer model, Llama 3.2 11B (Vision). We suggest considering it instead.

For more information, see comparison of Llama 3.2 11B (Vision) to other models and API provider benchmarks for Llama 3.2 11B (Vision).

Llama 3.1 Instruct 8B logo
Meta
•

Open weights model

•

Released July 2024

Llama 3.1 Instruct 8B API Provider Benchmarking & Analysis

Model Comparison

Analysis of API providers for Llama 3.1 Instruct 8B across performance metrics including latency (time to first token), output speed (output tokens per second), price and others. API providers benchmarked include Groq, Novita, CoreWeave, Amazon Bedrock, Cloudflare, DeepInfra (Turbo, FP8), DeepInfra, and Scaleway.

Fastest

#1
GroqGroq
638.8 t/s
#2
CloudflareCloudflare
192.9 t/s
#3
AmazonAmazon
181.8 t/s
#4
ScalewayScaleway
141.4 t/s
#5
NovitaNovita
140.8 t/s

Output speed

Total 8 providers

Lowest Latency

#1
CloudflareCloudflare
0.70 s
#2
AmazonAmazon
0.74 s
#3
NovitaNovita
0.85 s
#4
GroqGroq
0.97 s
#5
ScalewayScaleway
0.97 s

Time to first token

Total 8 providers

Lowest Price

#1
CloudflareCloudflare
$0.00
#2
DeepInfra (Turbo, FP8)DeepInfra (Turbo, FP8)
$0.02
#3
NovitaNovita
$0.02
#4
DeepInfraDeepInfra
$0.02
#5
GroqGroq
$0.05

Blended price (per 1M tokens)

Total 8 providers

Llama 3.1 8B is available through 8 API providers, each offering different performance characteristics and pricing. Below is a comparison of the key metrics across providers.

  • For output speed, the top providers are Groq (638.8 t/s), Cloudflare (192.9 t/s), and Amazon (181.8 t/s). Speed varies significantly across providers, with a 4015% difference between the fastest and slowest.
  • For latency, Cloudflare (0.70s), Amazon (0.74s), and Novita (0.85s) offer the lowest time to first token.
  • For pricing, Cloudflare (0.00), DeepInfra (Turbo, FP8) (0.02), and Novita (0.02) offer the lowest blended prices per 1M tokens.
  • Cloudflare offers both the lowest latency and best pricing, making it attractive for cost-conscious applications. Groq is the fastest option for throughput-intensive workloads.

Highlights

Speed

Output tokens per second · Higher is better

End-to-End Response Time

Seconds · Lower is better

Price

USD per 1M tokens (blended) · Lower is better

Pricing

Pricing: Cache Hit, Input, and Output

Price (USD per M Tokens) · Lower is better · 10,000 input tokens

Price per token for cached prompts (previously processed), typically offering a significant discount compared to regular input price, represented as USD per million tokens. The values shown here are the cache hit price; cache write and cache storage are billed separately and vary by provider — see "Cache pricing by provider" for detail.

Pricing: Blended Price

Blended at 7:2:1 (cache-input-output) · USD per 1M tokens (blended) · Lower is better

Price per token, shown in USD per million tokens. Price is a blend of cache hit, input, and output token prices using the selected ratio (default 7:2:1 cache-input-output).

Output Speed vs. Price

Blended at 7:2:1 (cache-input-output) · Output speed: output tokens per second · USD per 1M tokens (blended)
Most attractive quadrant
Pareto line

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Speed

Measured by Output Speed (tokens per second)

Output Speed: Llama 3.1 Instruct 8B

Output speed: output tokens per second · 10,000 input tokens

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API (e.g. OpenAI for o1) or the median across providers where a first-party API is not available (e.g. Meta's Llama models).

Latency vs. Output Speed

Latency: seconds to first token received · Output speed: output tokens per second · 10,000 input tokens
Most attractive quadrant
Pareto line

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Latency

Measured by Time (seconds) to First Token

Time to First Token

Seconds to first token received · Lower is better · 10,000 input tokens

Time to first token received, in seconds, after API request sent. For reasoning models which share reasoning tokens, this will be the first reasoning token. For models which do not support streaming, this represents time to receive the completion.

Figures represent median (P50) measurement over the past 72 hours to reflect sustained changes in performance.

End-to-End Response Time

Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed

End-to-End Response Time

Seconds to output 500 tokens, including reasoning model 'thinking' time · Lower is better · 10,000 input tokens

Seconds to receive a 500 token response. Key components:

  • Input time: Time to receive the first response token
  • Thinking time (only for reasoning models): Time reasoning models spend outputting tokens to reason prior to providing an answer. Amount of tokens based on the average reasoning tokens across a diverse set of 60 prompts (methodology details).
  • Answer time: Time to generate 500 output tokens, based on output speed

Key Comparison Metrics & API Features

Groq logoGroq
131k
Open
--
634
0.99
1.78
--
Novita logoNovita
16.4k
Open
--
130
0.85
4.69
--
CoreWeave logoCoreWeave
128k
Open
--
130
1.00
4.86
--
Amazon Bedrock logoAmazon Bedrock
128k
Open
--
172
0.73
3.64
--
Cloudflare logoCloudflare
30k
Open
--
185
0.70
3.40
--
FriendliAI logoFriendliAI
128k
Open
--
199
0.76
3.27
--
DeepInfra logoDeepInfra
131k
Open
--
17
1.11
30.42
--
DeepInfra logoDeepInfra
131k
Open
--
17
1.12
30.95
--
Scaleway logoScaleway
128k
Open
--
141
0.99
4.53
--

Frequently Asked Questions

Common questions about Llama 3.1 Instruct 8B providers

Llama 3.1 Instruct 8B is available through 8 API providers: Groq, Novita, CoreWeave, Amazon, Cloudflare, DeepInfra (Turbo, FP8), DeepInfra, and Scaleway. Each provider offers different performance characteristics and pricing.

Llama 3.1 Instruct 8B is currently available through 8 API providers that we benchmark and track.

The fastest providers for Llama 3.1 Instruct 8B by output speed are Groq (638.8 t/s), Cloudflare (192.9 t/s), and Amazon (181.8 t/s). Output speed measures how quickly tokens are generated after the model starts responding.

The providers with the lowest time to first token for Llama 3.1 Instruct 8B are Cloudflare (0.70s), Amazon (0.74s), and Novita (0.85s). Lower latency means faster initial response time.

The most affordable providers for Llama 3.1 Instruct 8B by blended price are Cloudflare ($0.00 per 1M tokens), DeepInfra (Turbo, FP8) ($0.02 per 1M tokens), and Novita ($0.02 per 1M tokens). Blended price uses a 7:2:1 cache hit/input/output token ratio.

The providers with the lowest input token pricing for Llama 3.1 Instruct 8B are Cloudflare ($0.00 per 1M input tokens), Novita ($0.02 per 1M input tokens), and DeepInfra (Turbo, FP8) ($0.02 per 1M input tokens).

The providers with the lowest output token pricing for Llama 3.1 Instruct 8B are Cloudflare ($0.00 per 1M output tokens), DeepInfra (Turbo, FP8) ($0.04 per 1M output tokens), and Novita ($0.05 per 1M output tokens).

Output speed for Llama 3.1 Instruct 8B varies significantly across providers. Groq is the fastest at 638.8 t/s, which is 41.2x faster than DeepInfra (Turbo, FP8) at 15.5 t/s.

7 of 8 providers support JSON mode for Llama 3.1 Instruct 8B: Groq, Novita, CoreWeave, Cloudflare, DeepInfra (Turbo, FP8), DeepInfra, and Scaleway.

7 of 8 providers support function calling for Llama 3.1 Instruct 8B: Groq, CoreWeave, Amazon, Cloudflare, DeepInfra (Turbo, FP8), DeepInfra, and Scaleway.

The best provider for Llama 3.1 Instruct 8B depends on your priorities: Groq offers the highest output speed, Cloudflare has the lowest latency, and Cloudflare provides the most competitive pricing.

When choosing a provider for Llama 3.1 Instruct 8B, consider: output speed (for throughput-intensive tasks), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and API features like JSON mode or function calling.

Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.

For information about Llama 3.1 Instruct 8B's intelligence, capabilities, modalities, and how it compares to other models, see the model overview page. View model overview