Reference — Models

The right model for the job.

The gap between open-weight and frontier models has nearly closed. The question is no longer "can open-weight compete?" It's "which model fits your workload, your compliance requirements, and your budget."

Quick pick

Start here.

Maximum security

Mistral Large 3

Apache 2.0, on-premises. No license restrictions. Full control.

Maximum intelligence

Claude Fable 5

Via API. Adaptive reasoning, 1M-token context. Pay per use.

Both — near-frontier, full control

GLM-5.2

Hosted dedicated. Leading open-weight model, MIT license. You own the weights.

Closed models

Frontier models (API-only).

Most intelligent. Least control. Pay per token.

Model	Provider	Best for	Context
Claude Fable 5	Anthropic	Deepest reasoning, complex coding, adaptive thinking	1M
GPT-5.5	OpenAI	General purpose, math, ecosystem	1M
Gemini 3.5 Pro	Google	Multimodal, video, longest preview context	2M (preview)
Grok 4 Fast	xAI	Largest practical context in production	2M

Pricing changes monthly — we quote current rates in every proposal

Trade-off:Maximum intelligence. Your data transits the vendor's infrastructure on every API call. The enterprise agreements we configure prevent training on your data.

Open models

Open-weight models (self-hostable).

Full control. Self-host or deploy on dedicated infrastructure.

Model	Parameters	License	Best for
DeepSeek V4 Pro	1.6T (MoE)	MIT	Reasoning & coding leader; lowest cost floor
GLM-5.2	754B (MoE)	MIT	Best overall open-weight; leads open intelligence indexes
Kimi K2.6	~1T (MoE)	Modified MIT	Agentic coding, long-horizon tool use (256K context)
Qwen 3.5	235B–397B (MoE)	Apache 2.0	Broad benchmark strength, multilingual, efficient
Mistral Large 3	675B (MoE)	Apache 2.0	Unrestricted license, multilingual
Llama 4	MoE family	Meta Llama	Long context (up to 10M), Meta ecosystem

Trade-off: 5–10% behind frontier on raw benchmarks. Zero data leakage. Full audit trail. You own the model. In 2024, the gap was 30%. Today, for most business tasks, open-weight is production-ready.

Legal

Licenses explained.

License	Commercial use?	Can modify?	Must share changes?
MIT (DeepSeek, GLM)	Yes	Yes	No
Apache 2.0 (Qwen, Mistral)	Yes	Yes	No
Modified MIT (Kimi)	Yes (UI attribution above 100M MAU / $20M rev)	Yes	No
Meta Llama	Yes (with limits)	Yes	No
Proprietary	Via API only	No	N/A

For regulated enterprises that need legal certainty, we recommend Apache 2.0 models (Mistral, Qwen) or MIT models (DeepSeek, GLM). No usage caps. No license surprises.

Framework

Intelligence vs. Control vs. Cost.

Intelligence
    ↑
    │  Frontier (API)
    │  ┌───────────────┐
    │  │ Claude Fable 5 │  ← Most intelligent
    │  │ GPT-5.5        │
    │  │ Gemini 3.5     │
    │  └───────────────┘
    │  Open-weight (self-hosted)
    │  ┌───────────────┐
    │  │ GLM-5.2        │  ← Near-frontier
    │  │ DeepSeek V4    │
    │  │ Kimi K2.6      │
    │  │ Qwen 3.5       │
    │  │ Mistral L3     │
    │  │ Llama 4        │
    │  └───────────────┘
    └──────────────────────→ Control

More control = open-weight on your hardware. More intelligence = frontier via API. The best setup is often both: open-weight for volume, frontier for hard problems.

Model landscape last verified July 2026. This changes fast — new models are released weekly. Talk to us for current recommendations.

Contact

Not sure which model fits?

We'll map your use case to the right model — and the right deployment mode.

Get a model recommendation →