Model library / IQuestLab
IQuest-Q1 is a 320 billion-parameter text-only language model from IQuestLab aimed at agentic coding, reasoning and tool use.
Parameters in total
Context length in tokens, as published
Weights at the released precision
Licence declared on the repository
Overview
IQuest-Q1 is a 320.3 billion-parameter language model from IQuestLab, described as a mixture-of-experts architecture for agentic coding, reasoning and multi-step tool use, though it has no expert layers. The publisher positions it for code generation, tool-calling agents and complex reasoning tasks, and aims to assist developers in automated debugging and workflow orchestration.
Deploying IQuest-Q1 requires handling a 320.3 billion-parameter checkpoint stored in bfloat16 format. The model supports a 524,288-token context window, but the licence is listed as "other", so users must review the specific terms. Recommended serving stacks include SGLang or vLLM, with pre-built Docker images provided for production use.
Specification
| Repository | IQuestLab/IQuest-Q1 |
| Publisher | IQuestLab |
| Published | 28 September 2026 |
| Architecture | IQuestQ1ForCausalLM |
| Model type | iquest_q1 |
| Parameters | 320,318,615,552 |
| Layers | 88 |
| Hidden size | 3,072 |
| Attention heads | 48, 8 key/value heads |
| Context length | 524,288 tokens |
| Vocabulary | 160,000 tokens |
| Weight format | bfloat16 |
| Licence | other |
Memory
Parameter count multiplied by the bytes each format uses. The key-value cache comes on top.
| Precision | Weights | Notes |
|---|---|---|
| bf16 | 596.6 GiB | Full precision as released |
| fp8 | 298.3 GiB | 8-bit, near-lossless on most models |
| int4 | 149.2 GiB | 4-bit, smallest footprint |
The cache costs 352 KiB per token at 16-bit, so the full 524,288-token context of one request needs about 176.0 GiB. Concurrency multiplies that number, not the weights.
That figure is an upper bound: this model uses sliding-window attention with a 4,096-token window on some layers, so long requests cache less than this.
Hardware
At the released precision, with 90% of the card given to the server and 2.5 GiB kept for the runtime. The context column is what is left for the cache on that setup.
| GPU | Cards needed | Free for cache | Context it holds |
|---|---|---|---|
| RTX 4090 24GB | more than 8 | — | — |
| RTX 5090 32GB | more than 8 | — | — |
| L40S 48GB | more than 8 | — | — |
| A100 80GB | more than 8 | — | — |
| H100 80GB | more than 8 | — | — |
| RTX PRO 6000 96GB | 8 | 74.6 GiB | about 222,107 tokens |
| H200 141GB | 8 | 398.6 GiB | the full 524,288 tokens |
Serving
A vLLM launch line for the shape above. Check the model card for a runtime the publisher recommends.
vllm serve IQuestLab/IQuest-Q1 \ --tensor-parallel-size 8 \ --max-model-len 221184 \ --gpu-memory-utilization 0.90
Questions
The weights take about 597 GiB at the released precision. No single card in the table above holds that, so it needs several GPUs or a lower precision.
The repository declares other. Read the licence text before commercial use: the name alone does not tell you what is allowed.
The configuration allows 524,288 tokens. Whether the whole window is usable depends on the memory left for the cache, which the hardware table works out per card.
This page is a reference, and Aiqre does not serve this model today. We deploy open-weight models on dedicated EU hardware on request, so if you want an OpenAI-compatible endpoint for it, email hello@aiqre.com with the model name and roughly what volume you expect.
Also in the library
Ling-3.1-flash is a 572.1 billion-parameter hybrid reasoning language model with a 262,144-token context window.
Beam is an open-source, instruction-tuned language model from Reflection using a sparse mixture-of-experts architecture.
MiMo-V2.6-Flash-MOPD is a 310.8 B parameter sparse-MoE model that merges domain-specialised teachers to reduce tool-call repetiti…
MiMo-V2.6-Pro-RL is a 524-billion-parameter sparse MoE model designed for reinforcement-learning-driven self-improvement across t…
Hosting
Tell us the model and roughly what volume you expect. We reply with a price and an OpenAI-compatible endpoint, with your prompts never stored.