d3 27B has 26.896 billion total parameters. The selected Q4_K_M GGUF file is 19.404 GB.
Data snapshot: . 1 published test in this catalog.
Model file and details
Provider
vLLM Semantic Router
Total parameters
26,895,998,464
Exact file size
19,403,686,112 bytes
Quantization / precision
Q4_K_M
Weight repository
ggml-org/vllm-d3-GGUF
Original model repository
vllm-sr/d3
Weight revision
de9670136f5ddd83e677c89ce041d5b9bb3b0dcb
Model revision
093cffbaa7be6f5dfe14252ae7e80c396c06dbd9
Runtime notes
Confirm model architecture and quantization support in your runner.
Recorded release date
2026-10-10
Date evidence
The official Apache-2.0 source checkpoint was created on 10 October 2026; ggml-org published the pinned automatic Q4_K_M llama.cpp conversion on 12 October 2026.
These scores were compiled from the linked sources. These published scores were not independently tested here. Different prompts, precision, thinking budgets and tools can change results.
Decision Index 0.3.1 · Composite
64.7 / 100
Publisher-reported Decision Index 0.3.1 composite across public, same-skill and new-domain typed-decision tasks. This is not a general chat, generation-speed or independent benchmark.
Benchmark efficiency: 3.334 published score points per GB of selected weights. This ratio is not an inference-speed measure; check the test settings to see whether the score covers the selected quantization.
Test settings: Publisher evaluation with Jev Decision Index 0.3.1. The card reports a 64.7 composite, including 65.5 on the public suite, 62.8 on same-skill tests and 56.6 on new-domain tasks. The released source weights answered all 140,178 public requests through the typed-decision interface.
Source-checkpoint result, not an independent test or a validation of the selected Q4_K_M GGUF. d3 is a 27B-family multimodal decision/classification model that returns option probabilities rather than chat text. The 19,403,686,112-byte language file excludes the separate 629,247,360-byte Q8_0 vision projector, runtime overhead and context/KV memory. The GGUF metadata declares a 262,144-token architecture context field, but practical context depends on runtime and available memory. Publisher latency figures were measured on one AMD Instinct MI325X with the source runtime and must not be transferred to this GGUF or consumer hardware. The GGUF card requires llama.cpp System One support and points to pull request 30383. Apache-2.0 license.
Context and version evidence: The repository and file name identify the selected model. Revision hashes are shown above when recorded. Tested context length, full hardware configuration and run date are not consistently recorded in this snapshot. Only settings explicitly stated in each benchmark record are known.
The 19.404 GB figure covers model weights in this file. It excludes the context cache, runtime, operating system and any separate multimodal files. It is not a measured RAM requirement or a token-generation speed.
These complete configurations pass the initial weight-size check for the selected 19.40 GB file. Allow extra memory for context and software. Check the model’s runtime notes and your inference software before buying.
Listing evidence: 2026-09-25. Amazon.com base price visible in an owner-supplied screenshot reviewed 25 Sep 2026; capture time not supplied. Dated reference, not a live quote. Excludes card signup offers, coupons, tax and shipping.
Listing evidence: 2026-10-02. Apple US launch price for the base M5 Max Mac Studio, published 22 September 2026: USD 2,499. Dated manufacturer reference, not an Amazon offer; tax and shipping may differ.
Listing evidence: 2026-09-24. GMKtec, EVO-X2 / 64GB / 1TB, recorded 22 Sep 2026; exact ASIN capacity identified 24 Sep. This is a dated reference, not a current Amazon price; tax, shipping and bundle contents may differ.
Memory matches are estimates, not tested compatibility or speed. As an Amazon Associate I earn from qualifying purchases.
Confirm model architecture and quantization support in your runner. The selected weights use 19.404 GB; this excludes context cache, runtime buffers and other programs.
These are the first listed tiers that pass our reserve policy, not measured minimum requirements. System RAM and VRAM remain separate. Use the memory tool to change the reserve for your context and runner.