Last refreshed 2026-10-14. Next refresh: weekly.
Why use LFM2.5-24B-A2B on Vercel AI Gateway?
Vercel AI Gateway offers LFM2.5-24B-A2B with pay-as-you-go pricing at $0.05/1M input tokens. Vercel AI Gateway is a unified AI proxy providing a single OpenAI-compatible API endpoint to 275+ models from 25+ providers including Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, xAI, Alibaba, Amazon, ByteDance, Cohere, MiniMax, MoonshotAI, KwaiPilot, Black Forest Labs, Recraft, Voyage AI, NVIDIA, and more.
Compare LFM2.5-24B-A2B across 3 providers to find the best fit for your use caseSetup recipe
Python + curlpip install openaiexport AI_GATEWAY_API_KEY=...import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["AI_GATEWAY_API_KEY"],liquid/lfm2.5-24b-a2bRequest example
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["AI_GATEWAY_API_KEY"],
base_url="https://ai-gateway.vercel.sh/v1"
)
response = client.chat.completions.create(
model="liquid/lfm2.5-24b-a2b",
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)Gotchas
- Use provider model ID "liquid/lfm2.5-24b-a2b", not the LLMReference slug "lfm-2.5-24b-a2b".
- creator/model-name e.g. kwaipilot/kat-coder-pro-v2
- The examples expect AI_GATEWAY_API_KEY; rename it only if your application config maps the new variable.
Compare LFM2.5-24B-A2B Across Providers
| Provider | Input (per 1M) | Output (per 1M) |
|---|---|---|
| Liquid AI | $0.10 | $1.00 |
| OpenRouter | $0.05 | $0.70 |
| Vercel AI Gateway | $0.05 | $0.70 |
Pricing
| Type | Price (per 1M) |
|---|---|
| Input tokens | $0.05 |
| Output tokens | $0.70 |
Capabilities
About LFM2.5-24B-A2B
LFM2.5-24B-A2B is Liquid AI's largest LFM2.5 model: a 24B-parameter mixture-of-experts reasoning model with about 2B active parameters per token, built for fast on-device agents. It keeps the LFM2 hybrid of gated short convolutions and GQA (40 layers, 64 experts, top-4 routing), extends context from 32,768 to 131,072 tokens, doubles the vocabulary to 128K, and scales pre-training from 17T to 33T tokens. Unlike LFM2-24B-A2B, it is reasoning-only (explicit chain of thought before the final answer) and is post-trained for tool calling and multi-step agent harnesses.