Last refreshed 2026-10-21. Next refresh: weekly.
Why use mammoth-1 on Vercel AI Gateway?
Vercel AI Gateway offers mammoth-1. 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 mammoth-1 across 2 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"],baseten/mammoth-1Request 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="baseten/mammoth-1",
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)Gotchas
- Use provider model ID "baseten/mammoth-1", not the LLMReference slug "mammoth-1".
- 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 mammoth-1 Across Providers
| Provider | Input (per 1M) | Output (per 1M) |
|---|---|---|
| Baseten API | — | — |
| Vercel AI Gateway | — | — |
Capabilities
About mammoth-1
mammoth-1 is a 27B open-weight decision model from Base Labs, Baseten's research lab, released October 21, 2026. It is a full fine-tune of Qwen3.5-27B that returns a probability for each predefined option instead of generating text, answering choice, yes/no and score questions for classification, routing and scoring. Block-bidirectional attention in the full-attention layers lets each question see its options, and one shared-prefix prefill serves every question in a request. It was trained on 6.21M decision examples with text-only input. Weights are on Hugging Face at baselabs/mammoth-1 under MIT.