Skip to content
RankShark

LFM3-1.7B

A compact 1.7B-parameter text model with a 256K-token context, post-trained for agents, tool calling, and structured outputs.

ModelOpen sourceby LiquidAIsince 2026
56.0

SharkScore

#29 in Models

Why it scores 56.0

LFM3-1.7B has a SharkScore of 56.0/100, ranked #29 of 608 in Models (updated 2026-11-19).

Each part compares LFM3-1.7B with other listings in Models. Evidence covers 87% of the score's weight.

Adoption
18 · 30%
User reviews
no data · 25%
Maintenance
98 · 20%
Trust & transparency
80 · 15%
Momentum
no data · 10%

Paying never changes this score. How ranking works · method v2026-11-18.2

RankShark review of LFM3-1.7B

Automated evaluation by RankShark · 18 Nov 2026 · method v1.0 · How we review

We were not paid for this review.

As of 2026-11-18, LFM3-1.7B is a 1.73B-parameter text model for local agents; its publisher advises against agentic coding and complex multi-step reasoning. [E1]

LFM3-1.7B is a post-trained, text-only language model intended for local agents, tool calling, structured outputs, and domain-specific fine-tuning. The publisher says it is not recommended for agentic coding or complex multi-step reasoning. [E1]

Best for

  • Developers building local agents or specialized applications. [E1] · model card
  • Applications that need tool calling, structured outputs, or domain-specific fine-tuning. [E1] · model card
  • Teams looking for a model with listed local CPU, GPU, and Apple Silicon inference options. [E1] · model card

Not for

  • Agentic coding or complex multi-step reasoning, which the publisher does not recommend for this model. [E1] · model card

Key features

  • Text model for agents. The vendor says it is a post-trained, text-only model designed for local agents, tool calling, and structured outputs. [E1] · model card
  • Long context. The vendor lists a 256K-token context length; RankShark measured 262,144 tokens on the OpenRouter listing. [E1; RankShark measured 2026-11-18] · model card
  • Tool use and tool search. The vendor says the model supports function-schema tool calling and deferred tools, with the application responsible for executing calls. [E1] · model card
  • Local inference variants. The vendor lists variants for vLLM, SGLang, Transformers, llama.cpp, Ollama, LM Studio, MLX, ONNX Runtime, and OpenVINO. [E1] · model card
  • Fine-tuning. The vendor says the model can be adapted to domains, formats, and tools through single-GPU fine-tuning. [E1] · model card

Pros

  • The publisher lists multiple local inference formats and runtimes, including CPU and GPU options through GGUF. [E1] · model card
  • The model supports tool calling and deferred tool loading according to its model documentation. [E1] · model card
  • The OpenRouter listing has per-token pricing and a measured context window of 262,144 tokens. [E2; RankShark measured 2026-11-18] · OpenRouter

Cons

  • The publisher says the model is not recommended for agentic coding or complex multi-step reasoning. [E1] · model card
  • The OpenRouter listing says prompts and outputs may be retained and used, but the available listing text does not specify the details. [E2] · OpenRouter
  • No free plan or free trial stated on the pages we checked.

Pricing

PlanPriceLimits
OpenRouter model listing (per-token pricing)$0.08 per million input tokens; $0.20 per million output tokens262,144-token context window; maximum completion of 8,192 tokens.

Checked 18 Nov 2026

Free plan
Not stated (checked 18 Nov 2026)
Free trial
Not stated (checked 18 Nov 2026)
Trains on customer data
Unclear: the pages disagree · OpenRouter
Data retention
Unclear: the pages disagree · OpenRouter
Data regions
Not stated (checked 18 Nov 2026)
Data processing agreement
Not stated (checked 18 Nov 2026)
AI models used
LFM3-1.7B: a post-trained, text-only language model with 1.73B parameters. [E1] · model card
API
The OpenRouter listing includes chat completions, completions, messages, and responses endpoints. [E2] · OpenRouter
SDKs
Not stated (checked 18 Nov 2026)
Webhooks
Not stated (checked 18 Nov 2026)
Platforms
The vendor lists inference support across GPU frameworks and local runtimes, including vLLM, SGLang, Transformers, llama.cpp, Ollama, LM Studio, MLX, ONNX Runtime, and OpenVINO. [E1] · model card
Security claims
No security certification stated (checked 18 Nov 2026)

What we measured

Context window (OpenRouter listing) · now
262,144 tokens
Price per million tokens (OpenRouter listing) · now
$0.08 in, $0.2 out
Package downloads · 30 days
109,972
How we tested · sources
  • Hourly uptime and response-time checks of the product's URL.
  • Repository, release and download data from GitHub, npm, PyPI and Hugging Face where they exist.
  • We did not sign up, pay, or use the product as a customer.
  • 18 Nov 2026: Plans or prices changed on the pricing page.
  • 18 Nov 2026: Plans or prices changed on the pricing page.
  • 18 Nov 2026: The data-use wording changed.

Something wrong? Report an error with a public link that shows it; corrections re-run the review.

Top alternatives to LFM3-1.7B

See all
  • 67.1

    Qwen3.8 27B is an open-weight dense vision-language model from Qwen.

    ModelOpen source
  • 65.5

    Qwen3.6-35B-A3B is an open-weight multimodal model from Alibaba Cloud with 35 billion total parameters and 3 billion active parameters per t

    ModelOpen source
  • 64.7

    Qwen3.8 Flash is a multimodal reasoning model from Alibaba.

    ModelOpen source
  • 64.3

    Qwen3.6 27B is a dense 27-billion-parameter language model from the Qwen Team at Alibaba, released in April 2026.

    ModelOpen source