Downloads 30d

20.5M

Hub likes

1.9K

Params

752M

The original GPT-2 checkpoint. Still the most downloaded causal language model on the Hub and a standard baseline.

Downloads 30d

32.2M

Hub likes

3.9K

Params

137M

Mid-size dense model from the Qwen3 family, with switchable thinking and non-thinking inference modes.

Downloads 30d

6.8M

Hub likes

2.2K

Params

8.2B

Downloads 30d

488.4K

Hub likes

1.2K

Params

31B

Instruction-tuned 7B chat model with strong multilingual coverage and a permissive licence.

Downloads 30d

6.1M

Hub likes

2.5K

Params

7.6B

Downloads 30d

4M

Hub likes

677

Params

18.7B

Downloads 30d

4.5M

Hub likes

740

Params

494M

Smallest checkpoint in Meta’s Open Pre-trained Transformer suite, released to reproduce GPT-3 class training openly.

Downloads 30d

4.7M

Hub likes

333

Params

125M

Downloads 30d

748.2K

Hub likes

32

Params

31.3B

Small instruction-tuned model that runs comfortably on modest hardware.

Downloads 30d

5.1M

Hub likes

937

Params

1.5B

Small Qwen3 model supporting both thinking and direct-answer modes in a single checkpoint.

Downloads 30d

3.6M

Hub likes

794

Params

4B

Compact instruction-tuned model built for on-device assistants and summarisation at low latency.

Downloads 30d

1.3M

Hub likes

2K

Params

1.2B

OpenAI’s smaller open-weight mixture-of-experts release, roughly 21B total parameters with about 3.6B active per token.

Downloads 30d

5.8M

Hub likes

5.3K

Params

20.9B

The most-liked chat model on the Hub. 128K context, instruction-tuned, widely used as a fine-tuning base.

Downloads 30d

4.5M

Hub likes

8.9K

Params

8B

Downloads 30d

2.5M

Hub likes

638

Params

3.1B

Downloads 30d

5.8M

Hub likes

661

Params

9B

The larger open-weight OpenAI release, around 117B total parameters with roughly 5.1B active per token.

Downloads 30d

2.4M

Hub likes

5.5K

Params

116.8B

Largest dense model in the Qwen3 line, aimed at reasoning and long-context work under a permissive licence.

Downloads 30d

3M

Hub likes

792

Params

32.8B

Downloads 30d

4.8M

Hub likes

90

Params

34.4B

Downloads 30d

2.7M

Hub likes

4.1K

Params

304.2B

Downloads 30d

4.2M

Hub likes

598

Params

35B

Downloads 30d

2.6M

Hub likes

1.1K

Params

4B

Downloads 30d

2.4K

Hub likes

370

Params

72.3B

Lightweight Qwen3 checkpoint aimed at latency-sensitive chat deployments.

Downloads 30d

3.6M

Hub likes

616

Params

2B

Downloads 30d

1.6M

Hub likes

63

Params

7.6B

Downloads 30d

2.4M

Hub likes

699

Params

9B

Part of a model suite trained on identical data in identical order, built specifically for interpretability research.

Downloads 30d

3.5M

Hub likes

66

Params

213M

Downloads 30d

1.9M

Hub likes

141

Params

4B

Downloads 30d

258.4K

Hub likes

81

Params

2.2B

Instruction-tuned small model from the Gemma 3 family, built from the same research as Gemini.

Downloads 30d

212.2K

Hub likes

1.3K

Params

1B

Code-specialised instruction model covering generation, completion and repair across many languages.

Downloads 30d

1.4M

Hub likes

868

Params

7.6B

Downloads 30d

2.5M

Hub likes

2K

Params

27B

Downloads 30d

1.8M

Hub likes

89

Params

14.8B

Mid-size instruction-tuned chat model, a common step up when 7B is not quite capable enough.

Downloads 30d

664.7K

Hub likes

383

Params

14.8B

Downloads 30d

767K

Hub likes

377

Params

32.8B

Downloads 30d

7.9M

Hub likes

1.6K

Params

685.4B

Downloads 30d

1.5M

Hub likes

1.1K

Params

35B

Very small language model trained on a curated corpus, intended for on-device and edge deployment.

Downloads 30d

2.6M

Hub likes

268

Params

135M

Downloads 30d

1.5M

Hub likes

519

Params

1M

Downloads 30d

1.5M

Hub likes

667

Params

88M

Downloads 30d

1.1M

Hub likes

53

Params

14.8B

Downloads 30d

481.6K

Hub likes

497

Downloads 30d

1.1M

Hub likes

868

Params

27B

Downloads 30d

1.1M

Hub likes

1.9K

Params

31.2B

Downloads 30d

1.8M

Hub likes

169

Params

14.4B

Mid-size member of the Qwen2.5-Coder family, trained on a large code-heavy corpus.

Downloads 30d

1.9M

Hub likes

213

Params

14.8B

Mixture-of-experts model with roughly 30B total parameters and about 3B active, trading memory for inference speed.

Downloads 30d

1.1M

Hub likes

982

Params

30.5B

Downloads 30d

857K

Hub likes

3K

Params

3.2B

Downloads 30d

1.4M

Hub likes

3.3K

Params

7.2B

Downloads 30d

1.6M

Hub likes

70

Params

8.2B

Downloads 30d

1.2M

Hub likes

2.4K

Params

290.9B

Downloads 30d

1.1M

Hub likes

587

Params

20.9B

Downloads 30d

1.2M

Hub likes

493

Params

494M

Downloads 30d

770

Hub likes

0

Downloads 30d

12.9M

Hub likes

536

Params

14.8B

Downloads 30d

635.7K

Hub likes

98

Params

1.7B

Downloads 30d

443.5K

Hub likes

118

Params

4B

Downloads 30d

344.9K

Hub likes

212

Params

79.7B

Downloads 30d

968.3K

Hub likes

50

Params

32.8B

Downloads 30d

852.6K

Hub likes

471

Params

135M