Ctrl K
Back to models

Meta

Llama 3.3 Instruct 70B

  • Released
  • Gated weights
  • Llama 3.3 Community License Agreement
Compare
Intelligence Index

Standard setting

7.7
Estimated · Artificial Analysis v4.3.2
API price from
$0.71 in / $0.71 out / 1M tokOpenRouter · CoreWeave · FP16
Context window
128K tok

Performance profile

ReasoningRank #95 of 107
49.8%
GPQA Diamond

Capabilities

Context window

Modalities

Measured setting

Integration

Reasoning trade-offs

Source (opens in a new tab)
Intelligence Index v4.3.2
StandardEstimated
Intelligence Index
7.7

Pricing & access

API route
OpenRouter · CoreWeave · FP16API
Tier & contextStandard

USD per 1M tokens

Input
$0.71
Output
$0.71
Cached input
$0.71

Run locally

GGUF
Context
Estimated memory

~46.1 GiB

RAM + VRAM

Estimate assumptions

One text chat at the selected context, with a 16-bit cache and app allowance. Operating system and image processing are extra; optimized caches can use less.

64 GiB RAMwithout GPU offload

Memory budgetOne text chat
Weights
39.6 GiB
Context cache
2.5 GiB16-bit cache
App allowance
4 GiB
Download
42.5 GB
Disk space
~52 GB
Available software
Ollama (opens in a new tab)LM Studio (opens in a new tab)Jan (opens in a new tab)KoboldCpp (opens in a new tab)llama.cpp (opens in a new tab)
Get GGUF (opens in a new tab)
Set up in Ollama · Q4_K_M · 8K context

Save the following as Modelfile. The context matches the estimate above.

FROM hf.co/lmstudio-community/Llama-3.3-70B-Instruct-GGUF:Q4_K_M
PARAMETER num_ctx 8192

Then run these commands in your terminal:

ollama create llmcurrent-local -f Modelfile
ollama run llmcurrent-local

Ollama import instructions (opens in a new tab) · For LM Studio or Jan, open the app’s model search, enter lmstudio-community/Llama-3.3-70B-Instruct-GGUF, choose Q4_K_M, and set context to 8,192 tokens in the app’s model settings. Software links above open each app’s setup guide.

vLLM setup (opens in a new tab)

Benchmarks

Full leaderboardSource (opens in a new tab)

Humanity’s Last Exam

Frontier reasoning

Llama 3.3 Instruct 70B · Standard#126 of 126

  1. 1Claude Opus 5.5Max61.4%
  2. 2Claude Fable 5.1Max59.1%
  3. 3Gemini 4 ArgonHigh57.1%
  5. 120Ministral 3 8BStandard4.3%
  6. 121Nova PremierStandard4.2%
  7. 122Command AStandard4.0%
  8. 123Granite 4.0 H SmallStandard3.8%
  9. 124Llama 4 ScoutStandard3.8%
  10. 125Phi-4Standard3.8%
  11. 126Llama 3.3 Instruct 70BStandard3.6%

Peers use their best measured result on this benchmark.