AI model library — official specs, cutoffs & comparisons

1,512 matching models

Mistral AIAvailable

Ministral-3-14B-Instruct-2512

Ministral · 25.12

API / Open weights

Context: 256,000 tokens
Max. output: Not documented
Knowledge cutoff: Not documented

Mistral AIAvailable

Ministral-3-3B-Instruct-2512

Ministral · 25.12

API / Open weights

Context: 256,000 tokens
Max. output: Not documented
Knowledge cutoff: Not documented

Mistral AIAvailable

Ministral-3-8B-Instruct-2512

Ministral · 25.12

API / Open weights

Context: 256,000 tokens
Max. output: Not documented
Knowledge cutoff: Not documented

Mistral AIAvailable

Mistral-Large-3-675B-Instruct-2512

Mistral · 25.12

API / Open weights

Context: 256,000 tokens
Max. output: Not documented
Knowledge cutoff: Not documented

AmazonAvailable

Nova 2 Lite

Nova · v1:0

API

Context: 1,000,000 tokens
Max. output: 65,536 tokens
Knowledge cutoff: Oct 2025Knowledge cutoff · month precision

AmazonAvailable

Nova 2 Sonic

Nova · v1:0

API

Context: 1,000,000 tokens
Max. output: 65,536 tokens
Knowledge cutoff: Not documented

AnthropicAvailable

Claude Opus 4.5

Claude Opus · claude-opus-4-5-20251101

API

Context: 200,000 tokens
Max. output: 64,000 tokens
Knowledge cutoff: May 2025Reliable knowledge cutoff · month precision

NVIDIAListed by publisher

NVIDIA-Nemotron-Parse-v1.1

Nemotron

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

NVIDIAListed by publisher

NVIDIA-Nemotron-Parse-v1.1-TC

Nemotron

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-1b

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-1b-base

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-350m

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-350m-base

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-h-1b

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-h-1b-base

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-h-350m

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-h-350m-base

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

NVIDIAListed by publisher

NVIDIA-Nemotron-Nano-12B-v2-VL-BF16

Nemotron

Open weights

Context: 128K tokens
Max. output: Not documented
Knowledge cutoff: Not documented

NVIDIAListed by publisher

Nemotron-H-4B-Base-8K

Nemotron

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

NVIDIAListed by publisher

Nemotron-H-4B-Instruct-128K

Nemotron

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

AnthropicAvailable

Claude Haiku 4.5

Claude Haiku · claude-haiku-4-5-20251001

API

Context: 200,000 tokens
Max. output: 64,000 tokens
Knowledge cutoff: Feb 2025Reliable knowledge cutoff · month precision

IBMListed by publisher

granite-4.0-h-micro

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-h-micro-base

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

IBMListed by publisher

granite-4.0-h-small

Granite

Open weights

Context: Not documented
Max. output: Not documented
Knowledge cutoff: Not documented

Go beyond the spec sheet

The decisions behind the numbers.

Source-checked guides to pricing, usable context, and access restrictions.

Claude Sonnet 4.6 pricing: API costs, caching and real workload math GPT-5.4 context window: the 1.05M limit, output budget and cost threshold Gemini 2.5 Pro context window: input limits, access and long-document costs GPT-5.4 mini context window: 400K capacity and practical task budgets