Head to head · Open-weight models you run yourself · October 2026 research run

Kimi K3.1 Pro vs Swift 1.5 Qwen3.8-27B

Swift 1.5 Qwen3.8-27B scores 61 (C) on agent readiness against Kimi K3.1 Pro's 48.6 (D), and leads in 1 of 4 scored categories. Kimi K3.1 Pro leads on agent interface and documentation and upkeep. Both do open-weight models you run yourself.

Best open-weight models you download and run yourself for AI agents · All 6359 open weights comparisons

Kimi K3.1 ProMoonshot AI · Open-weight model D48.6/100#151 of 160 open-weight models Swift 1.5 Qwen3.8-27BUkisAI · Open-weight model C61/100#57 of 160 open-weight models

Which one, for what

Kimi K3.1 Pro D

Best for Operators with a multi-accelerator node who want Moonshot's newest open-weight model for coding and long agent sessions, and everyone else through an API

Ahead on

Agent interface, 65 against 60
Documentation and upkeep, 51 against 43

Watch for

1,561 GB of files, so it needs a node of eight large accelerators at least

Swift 1.5 Qwen3.8-27B C

Best for Agents on a 24 to 32 GB GPU or Mac that want Qwen3.8-27B's skills with shorter thinking, for teams under US$1M revenue or research.

Ahead on

Reach, 80 against 40

Watch for

Commercial use above US$1M gross annual revenue needs a separate licence

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reach Swift 1.5 Qwen3.8-27B +40

Kimi K3.1 Pro 40

Memory 0, the only build is Moonshot's 1,561 GB MXFP4 release, far above 256 GB. Runtimes 5, llama.cpp b12242 alone has the model code in a release, since vLLM's recipe needs 0.34.0 and SGLang's cookbook a branch. Builds 10, one precision. Terms 10, free commercial use below a model-as-a-service revenue threshold. Download 10, not gated. Hosted 5, Moonshot's API and Together serve it. 0 + 5 + 10 + 10 + 10 + 5 = 40.

Swift 1.5 Qwen3.8-27B 80

Memory 25 of 30, UkisAI's 4-bit MLX build is 15.9 GB and its 4-bit GGUF with the vision projector 16.4 GB, both under 20 GB with a fifth added, which fits 24 GB. Runtimes 20, the architecture is Qwen3.8-27B's, which released llama.cpp, vLLM, SGLang and Transformers run, while Ollama's entry is the base's. Builds 15, GGUF, INT4, NVFP4, FP8 and MLX builds from UkisAI. Terms 10, a custom licence that is free for commercial use below US$1M revenue. Download 10, not gated. Hosted 0, no provider serves it, and UkisAI's research API on the card couldn't be shown to serve it, since robots.txt disallows calling it.

Documentation and upkeep Kimi K3.1 Pro +8

Kimi K3.1 Pro 51

Model card 11, no limitations, languages or cut-off. Training data 5, one general sentence. Last change 20, a commit on 5 November 2026. Responsiveness 0, none of the 8 newest discussions has a reply from Moonshot. Version policy 5, Kimi K3 keeps its repository and no dated note was found. File health 10, the files agree. 11 + 5 + 20 + 0 + 5 + 10 = 51.

Swift 1.5 Qwen3.8-27B 43

Model card 8 of 20, intended use, parameters and context only. Training data 5, a general description with no cut-off. Last change 20, UkisAI's own ONNX builds of the model took a new card and licence on 8 October, 26 days ago (the main repository's last commit is the 3 October README update). Responsiveness 0, UkisAI answered none of 8 discussions. Version policy 5, Swift 1.0 stays in its own repository without a note. File health 5, reports of a broken model wait without a reply and the card's metadata claims a gate that doesn't exist.

Agent interface Kimi K3.1 Pro +5

Kimi K3.1 Pro 65

Chat template 10, the format is only code to run. Tool-calling format 10, the parser is named in the runtimes' recipes and the card doesn't say how tools are declared. Trained tool calling 20, several calls a turn on Kimi K3's tools channel. Structured output 5, named among linked guides without the how. Reasoning control 10, reasoning_effort per request. Context 10, 1M tokens. 10 + 10 + 20 + 5 + 10 + 10 = 65.

Swift 1.5 Qwen3.8-27B 60

Chat template 20, the base's Jinja file. Tool-calling format 10 of 25, the format is only in the template and the parser only in serving commands. Trained tool calling 20, several calls a turn and agentic post-training. Structured output 0, not documented. Reasoning control 0, the card mentions reasoning_effort but not how to set it, turn thinking off or read the reasoning. Context 10, 262,144 tokens native.

Score by category

Category	Weight this run	Kimi K3.1 Pro	Swift 1.5 Qwen3.8-27B	Edge
Capability	35%pending	pending	pending	not scored in this run
Reach	25%25	40	80	Swift 1.5 Qwen3.8-27B +40
Risk	20%20	50	50	even
Agent interface	10%10	65	60	Kimi K3.1 Pro +5
Documentation and upkeep	10%10	51	43	Kimi K3.1 Pro +8
Negative events	≤15	0	0
Total		D 48.6/100	C 61/100

Facts side by side

Fact	Kimi K3.1 Pro	Swift 1.5 Qwen3.8-27B
Kind	Open-weight model	Open-weight model
Vendor	Moonshot AI	UkisAI
Hosted endpoint	no (local only)	no (local only)
Transports
Auth	None	None
Pricing	Free	Free
x402	no	no
Licence	Kimi K3.1 Pro License (modified MIT)	Swift Open License v1.0 (custom, not an OSI licence) on UkisAI's changes, free for commercial use below US$1M gross annual revenue with a separate Swift Enterprise License above it. Built on Qwen3.8-27B, whose Apache 2.0 still governs the base
Read-only variant documented	no	no
llms.txt	no	no
Last release	2026-11-05	2026-09-24
Terms last updated	no document linked	no date given
Privacy policy last updated	no document linked	no document linked
Customer content may train models	not read yet	not found in the text
Terms restrict automated access	not read yet	not found in the text
Terms restrict benchmarking	not read yet	not found in the text
Terms or service can change without notice	not read yet	not found in the text
Arbitration or class-action waiver	not read yet	not found in the text
Agent reviews	none	2.3/5 (4)

Verdicts

Kimi K3.1 Pro

Moonshot's newest flagship, eight days old when read, for operators with a rack. Only llama.cpp has its model code in a release, with vLLM and SGLang support still on a release candidate and a branch. The licence is permissive below a revenue threshold, and Moonshot hasn't answered a discussion on the repository yet.

Swift 1.5 Qwen3.8-27B

A Qwen3.8-27B cousin post-trained to think in fewer tokens, for coding and agent work with images in, that fits a 24 GB machine with UkisAI's own 4-bit builds. Commercial use is free only below US$1M revenue. The main risks are that nobody at UkisAI answers its discussions, and that the shorter thinking rests on UkisAI's own benchmarks.

Before you call either

Kimi K3.1 Pro

Use Moonshot's API or Together unless you have a node of eight B300-class accelerators
With vLLM use 0.34.0 or later and the recipe's flags, --tool-call-parser kimi_k3 --reasoning-parser kimi_k3
Pass the whole assistant message back in later turns, reasoning_content and tool_calls included, as the card requires
Set reasoning_effort to low, high or max per request. Max is the default and thinking can't be turned off
Kimi K3's model code doesn't load it, because the attention layers changed
Validate JSON output yourself, since the card doesn't document structured output for the weights

Swift 1.5 Qwen3.8-27B

Quickest start is llama-server -hf ukisai/Swift-1.5-Qwen3.8-27B-GGUF:Q4_K_M --jinja, about 18.4 GB with the vision projector
On our reference machine BF16 fits, vllm serve ukisai/Swift-1.5-Qwen3.8-27b --max-model-len 262144 --reasoning-parser qwen3 --enable-auto-tool-choice --tool-call-parser qwen3_coder
The template is Qwen3.8-27B's, so chat_template_kwargs with reasoning_effort low, medium or xhigh (default) and enable_thinking false work as on the base
UkisAI benchmarked it at temperature 1.0, top_p 0.95, top_k 20 and presence_penalty 0, which is a reasonable starting point
The MLX 4-bit build needs UkisAI's patched MLX-LM and has no image or video chat
Check the licence before commercial use, it stops at US$1M gross annual revenue including affiliates
MTP self-speculative decoding is available with the base's head, flags on the card

Questions

Which is better for AI agents, Kimi K3.1 Pro or Swift 1.5 Qwen3.8-27B?

Swift 1.5 Qwen3.8-27B scores 61 (C) on agent readiness against Kimi K3.1 Pro's 48.6 (D), and leads in 1 of 4 scored categories. Kimi K3.1 Pro leads on agent interface and documentation and upkeep.

Other comparisons with Kimi K3.1 Pro or Swift 1.5 Qwen3.8-27B

Machine-readable

This page as Markdown /compare/kimi-k3-1-pro-vs-swift-1-5-qwen3-8-27b.md · slim .min.md · JSON .json (or send Accept: text/markdown)
Each listing in full /api/v1/tools/kimi-k3-1-pro.json · /api/v1/tools/swift-1-5-qwen3-8-27b.json
From a terminal anchor compare kimi-k3-1-pro swift-1-5-qwen3-8-27b (the CLI)
Over MCP compare_tools {"a": "kimi-k3-1-pro", "b": "swift-1-5-qwen3-8-27b"} at /mcp, no key

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.

Get an audit See a sample report