v2.2 · 2026-06-12
Dolphin 3.0 Llama 3.1 8B
dolphin-3-0-llama-3-1-8b
Dolphin 3.0 Llama 3.1 8B is a 8.03B text chat model profile for local AI planning. This page records provider, license, context, quantization, hardware-fit estimates, setup hints, source links, and caveats so readers do not choose a model by name alone.
PROVIDER
dphn / Cognitive Computations
FAMILY
dolphin
MODEL TYPE
text-chat
PARAMETERS
8.03B
MODALITIES
text
ARCHITECTURE
Llama 3.1 8B-derived decoder-only Transformer fine-tuned with ChatML formatting
CONTEXT WINDOW
128,000 tokens
TRAINING TOKENS
—
RELEASE DATE
—
GGUF REPOSITORIES
Publisher-hosted GGUF artifact with Q4_K_M listing and llama.cpp/Ollama HF commands.
bartowski/Dolphin3.0-Llama3.1-8B-GGUF community conversion
Community GGUF conversion; use only when a required quantization is present and file size is listed in the cited source.
SETUP HINTS
OLLAMA
ollama run dolphin3:8bLM STUDIO
Use dphn/Dolphin3.0-Llama3.1-8B-GGUF and select Q4_K_M; verify any higher quantization file size before cataloging it.
LLAMA.CPP
llama-server -hf dphn/Dolphin3.0-Llama3.1-8B-GGUF:Q4_K_MQUANTIZATION FILE SIZES (GGUF)
Q4_K_M
4.92 GB
Quantization note: Q4_K_M
RAM / VRAM ESTIMATES
MIN RAM
8 GB
COMFORTABLE RAM
16 GB
MIN VRAM
6 GB
COMFORTABLE VRAM
8 GB
These are conservative local-inference estimates, not official hardware requirements.
HARDWARE FIT
CPU only (no GPU)
yes
8 GB RAM
limited
16 GB RAM
comfortable
32 GB RAM
comfortable
8 GB VRAM
good
12 GB VRAM
comfortable
24 GB VRAM
comfortable
Apple Silicon (unified memory)
comfortable
Hardware fit values are conservative local-inference estimates based on GGUF size plus runtime and KV-cache overhead. Actual requirements depend on context length, quantization, runtime, GPU offload, and other running apps.
BEST FOR
·steerable local assistant workflows
·coding and math assistance in the 8B class
·function-calling and agentic local experiments
·users who want a less refusal-heavy local model
AVOID IF
·you need official Meta safety alignment
·you need a model with a clearly verified release date
·you need verified Q5_K_M or Q8_0 file sizes for the publisher GGUF artifact
CAVEATS
·RAM and VRAM estimates are conservative local-inference estimates based on GGUF file size plus runtime and KV-cache overhead, not official dphn requirements.
·The official/publisher model card and GGUF artifact are clear enough to include this target; no replacement was used.
·Release date was not confirmed in cited sources from the fetched model card, so releaseDate is null.
·Only the Q4_K_M file size was verified from the fetched publisher GGUF listing; q5FileSizeGb and q8FileSizeGb are null rather than inferred.
·This is a derivative of Llama 3.1 8B and remains subject to the Llama 3.1 license.
·The context-window value is inherited from the Llama 3.1 8B base family rather than independently stated as a Dolphin-specific training limit in the fetched model card.
SOURCE URLS
FIELD EVIDENCE
canonicalModelCardUrlhuggingface.co/dphn/Dolphin3.0-Llama3.1-8B
parameterSizeBollama.com/library/dolphin3%3A8b
architecturehuggingface.co/dphn/Dolphin3.0-Llama3.1-8B-GGUF
contextWindowTokenshuggingface.co/blog/llama31
supportsToolshuggingface.co/dphn/Dolphin3.0-Llama3.1-8B
q4FileSizeGbhuggingface.co/dphn/Dolphin3.0-Llama3.1-8B-GGUF
LANGUAGES
English
CAPABILITIES
tool callingcode tuned