SAVRN
Search Contact SAVRN

SAVRN Model Hub · Comparisons

Kimi-K3-DSpark vs Qwen3-1.7B

Kimi-K3-DSpark has 2.2B parameters and Qwen3-1.7B has 2B parameters; at 16-bit, Kimi-K3-DSpark needs about 5.4 GB (1x MI300X from $1.85 an hour) and Qwen3-1.7B about 4.9 GB (1x MI300X from $1.85 an hour).

Published metadata for 2 models, each read from its own repository.
Field Kimi-K3-DSpark
RadixArk/Kimi-K3-DSpark
Qwen3-1.7B
Qwen/Qwen3-1.7B
Publisher RadixArk Qwen
Task Text generation Text generation
Modality Text Text
Parameters, as reported 2.2B parameters 2B parameters
Architecture DSparkDraftModel Qwen3ForCausalLM
Library transformers transformers
Context length 1,048,576 tokens 40,960 tokens
Repository size 4.5 GB 4.1 GB
Artifact formats safetensors safetensors
License Not stated apache-2.0
Access Open weights, no gate Open weights, no gate
Memory at 16-bit (weights and margin) 5.4 GB 4.9 GB
Cheapest GPUs at 16-bit, per hour 1x MI300X, $1.85 1x MI300X, $1.85
Memory at 4-bit (weights and margin) 1.3 GB 1.2 GB
Cheapest GPUs at 4-bit, per hour 1x MI300X, $1.85 1x MI300X, $1.85
Revision viewed 3c5bac301d9c 70d244cc86cc
Downloads reported by the hub 3.5M 3.8M
Last observed 2026-09-18 2026-09-18

An evaluation row appears only where at least two of these models report the same benchmark with the same stated configuration, metric, unit and setup. Different evaluators stay named in each cell. Values are shown as reported: no unit conversion, no ranking.

SAVRN's Notes on Kimi-K3-DSpark

Read this as a companion, not a model you serve on its own. RadixArk trained Kimi-K3-DSpark as a DSpark speculator for the Kimi K3 target, a 2.2 billion parameter draft built for faster inference through speculative decoding. Five layers and 4.5 GB of weights keep it light: 5.4 GB of memory at 16-bit, 2.7 GB at 8-bit, 1.3 GB at 4-bit, and one MI300X with 192 GB at $1.85 an hour on-demand is the cheapest listing we track. That covers the draft only; the Kimi K3 target sizes your box.

The license field is empty, so get written terms from RadixArk before deployment; open access is not permission to run it commercially. Match your stack too: the checkpoint was trained with SpecForge on hidden states from a live SGLang target engine, and the 1,048,576 token context is the draft's ceiling, so the target needs the same window.

SAVRN's Notes on Qwen3-1.7B

Where does a text generator with 2B parameters belong in a facility we run? On a card that is already busy. At 4-bit it needs 1.2 GB of memory; at 16-bit, nothing quantized, 4.9 GB. The cheapest listed setup is one MI300X with 192 GB at $1.85 an hour on-demand, so we treat a footprint this small as a tenant, not a workload: it rides beside bigger jobs and adds nothing to the hardware bill.

Apache 2.0 allows commercial use, modification and redistribution; keep the license and copyright notices, state significant changes, and a fine-tune can ship to customers. Context is 40,960 tokens, what it holds in one pass. This checkpoint is derived from Qwen3-1.7B-Base, so a team planning its own instruction tuning starts from the base. No Index host sells it by the token, so the card price is the only price.

Questions

Which is larger, Kimi-K3-DSpark or Qwen3-1.7B?

Kimi-K3-DSpark (2.2B parameters) is larger than Qwen3-1.7B (2B parameters), by the parameter counts their publishers report.

Which is cheaper to run, Kimi-K3-DSpark or Qwen3-1.7B?

At 4-bit, Kimi-K3-DSpark fits on 1x MI300X from $1.85 an hour and Qwen3-1.7B on 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use Qwen3-1.7B commercially?

Yes. Qwen3-1.7B is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

Related Comparisons