Meta model product
Llama 3.2 3B Instruct is a large language model that supports a context length of 128K tokens and are state-of-the-art in their class for on-device use cases like summarization, instruction following, and rewriting tasks running locally at the edge.
Updated Aug 10, 2026. Default version: Llama 3.2 3B Instruct
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Llama 3.2 3B Instruct.
| InfiniteBench/En.MC | 0.6 | 1 | 1 | 100.0% | C | |
| InfiniteBench/En.QA | 0.2 | 1 | 1 | 100.0% | C | |
| NIH/Multi-needle | 0.8 | 1 | 1 | 100.0% | C | |
| Open-rewrite | 0.4 | 1 | 1 | 100.0% | C | |
| TLDR9+ (test) | 0.2 | 1 | 1 | 100.0% | C | |
| BFCL v2 | 0.7 | 4 | 5 | 25.0% | C | |
| Nexus | 0.3 | 4 | 4 | 0.0% | C | |
| ARC-C | 0.8 | 17 | 34 | 51.5% | C | |
| HellaSwag | 0.7 | 24 | 27 | 11.5% | C | |
| MGSM | 0.6 | 27 | 31 | 13.3% | C | |
| GSM8k | 0.8 | 40 | 48 | 17.0% | C | |
| IFEval | 0.8 | 56 | 65 | 14.1% | C | |
| MATH | 0.5 | 60 | 71 | 15.7% | C | |
| MMLU | 0.6 | 96 | 100 | 4.0% | C | |
| GPQA | 0.3 | 218 | 233 | 6.5% | C |
Preference and agent-evaluation signals from published Arena datasets.
| text | german | 273 | 1056.0 | 232 | N/A | |
| text style control | german | 273 | 1110.0 | 232 | N/A | |
| text | expert | 302 | 1089.4 | 499 | N/A | |
| text | industry medicine and healthcare | 306 | 1103.6 | 441 | N/A | |
| text style control | expert | 306 | 1176.3 | 499 | N/A | |
| text | english | 311 | 1179.8 | 4,157 | N/A | |
| text style control | industry medicine and healthcare | 311 | 1201.8 | 441 | N/A | |
| text | industry mathematical | 312 | 1128.5 | 980 | N/A | |
| text | industry legal and government | 313 | 1115.3 | 506 | N/A | |
| text | hard prompts english | 314 | 1142.7 | 1,377 | N/A | |
| text | math | 316 | 1125.9 | 1,136 | N/A | |
| text style control | industry mathematical | 316 | 1173.8 | 980 | N/A | |
| text style control | math | 318 | 1164.9 | 1,136 | N/A | |
| text style control | english | 320 | 1224.2 | 4,157 | N/A | |
| text | multi turn | 322 | 1105.1 | 1,299 | N/A | |
| text style control | hard prompts english | 323 | 1208.7 | 1,377 | N/A | |
| text | longer query | 324 | 1100.7 | 1,046 | N/A | |
| text | industry entertainment and sports and media | 325 | 1084.0 | 1,149 | N/A | |
| text | industry life and physical and social science | 325 | 1106.3 | 1,355 | N/A | |
| text style control | multi turn | 325 | 1148.7 | 1,299 | N/A | |
| text style control | industry entertainment and sports and media | 326 | 1141.8 | 1,149 | N/A | |
| text style control | industry legal and government | 326 | 1188.0 | 506 | N/A | |
| text | hard prompts | 327 | 1095.0 | 2,216 | N/A | |
| text | industry software and it services | 328 | 1104.3 | 2,165 | N/A | |
| text | instruction following | 328 | 1089.3 | 3,171 | N/A | |
| text style control | creative writing | 328 | 1144.3 | 1,070 | N/A | |
| text | creative writing | 329 | 1094.5 | 1,070 | N/A | |
| text | industry business and management and financial operations | 329 | 1073.4 | 1,012 | N/A | |
| text | overall | 330 | 1109.7 | 7,936 | N/A | |
| text | coding | 330 | 1098.2 | 1,351 | N/A | |
| text style control | chinese | 332 | 1128.5 | 721 | N/A | |
| text | chinese | 333 | 1016.3 | 721 | N/A | |
| text style control | longer query | 333 | 1158.6 | 1,046 | N/A | |
| text | exclude ties | 334 | 933.8 | 5,442 | N/A | |
| text style control | industry life and physical and social science | 334 | 1185.2 | 1,355 | N/A | |
| text | industry writing and literature and language | 335 | 1095.6 | 2,118 | N/A | |
| text style control | industry writing and literature and language | 335 | 1141.8 | 2,118 | N/A | |
| text style control | instruction following | 335 | 1145.5 | 3,171 | N/A | |
| text style control | industry business and management and financial operations | 337 | 1145.9 | 1,012 | N/A | |
| text style control | overall | 338 | 1166.4 | 7,936 | N/A | |
| text style control | hard prompts | 338 | 1167.6 | 2,216 | N/A | |
| text style control | exclude ties | 341 | 1008.8 | 5,442 | N/A | |
| text style control | industry software and it services | 343 | 1174.1 | 2,165 | N/A | |
| text | russian | 344 | 948.1 | 933 | N/A | |
| text style control | coding | 344 | 1176.0 | 1,351 | N/A | |
| text style control | russian | 345 | 1004.6 | 933 | N/A | |
| text style control | non english | 347 | 1087.6 | 3,779 | N/A | |
| text | non english | 348 | 1018.6 | 3,779 | N/A |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| Nvidia | meta/llama-3.2-3b-instruct | global | N/A | N/A | 32.8K | |
| Inference | meta/llama-3.2-3b-instruct | global | $0.02 | $0.02 | 16K | |
| LLM Gateway | llama-3.2-3b-instruct | global | $0.03 | $0.05 | 32.8K | |
| NovitaAI | meta-llama/llama-3.2-3b-instruct | global | $0.03 | $0.05 | 32.8K | |
| NanoGPT | meta-llama/llama-3.2-3b-instruct | global | $0.0306 | $0.0493 | 131.1K | |
| OpenRouter | meta-llama/llama-3.2-3b-instruct | global | $0.05 | $0.33 | 131.1K | |
| Cloudflare Workers AI | @cf/meta/llama-3.2-3b-instruct | global | $0.0509 | $0.335 | 80K | |
| Cloudflare AI Gateway | workers-ai/@cf/meta/llama-3.2-3b-instruct | global | $0.051 | $0.34 | 128K | |
| Pioneer | meta-llama/Llama-3.2-3B-Instruct | global | $0.10 | $0.335 | 131.1K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Llama 3.2 3B Instruct | 20.4 | 3.2B | 128K | 128K | No | Llama 3.2 Community License |
A concise description based on the published model registry.
Llama 3.2 3B Instruct is a large language model that supports a context length of 128K tokens and are state-of-the-art in their class for on-device use cases like summarization, instruction following, and rewriting tasks running locally at the edge.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Llama 3.2 3B.
Llama 3.2 3B's default version was released on Sep 25, 2024.
No official standard PAYG price is currently available for Llama 3.2 3B. The lowest tracked third-party offer starts at $0.02 input and $0.02 output via Inference.
Llama 3.2 3B is published under Meta in the model registry.
The default version has a 128K token context window.
No. The default version is not marked as having publicly available weights.
9 published provider offerings are linked to the default version.
Nearby ranked alternatives include Jamba 1.5 Mini, IBM Granite 4.0 Tiny, Qwen2.5 Omni 7B.