Up to date
Qwen Apache 2.0 out 2 years ago
14.8B parameters, text in, text out, 128K tokens of context. 1 provider serves it.
Verified 2026-10-25 Rank #295 in LLMs · score 84.1 method v1 Sources 3 Downloads 30d 85.2k · #330 of 619 tracked
Picked by rule from 34 tracked pieces: the lab's own post, the most shared, a field report, the latest.
The 3 picks above, plus these 22. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.
Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.
| # | Cheapest live row | In /1M | Out /1M | Context | Quant | Seen |
|---|---|---|---|---|---|---|
| 1 | Featherless | — | — | — | — | Oct 25 |
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
Served by 1 API provider; heavily run locally (78 quantized builds); a big base for builders (230 fine-tunes, adapters and merges).
Top Spaces running it: pliny-the-prompter/obliteratus ♥ 621 · ISEEKYAN/megatron_memory_estimator ♥ 35 · ISEEKYAN/megatron_memory_estimator_old ♥ 19 · artificialguybr/CanIRunThisLLM ♥ 9
Weights only, an estimate: leave room for the KV cache. Method
About 8.9 GB at INT4, 18 GB at FP8, 35 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
Weights, every dated article, and repo commits on one rail.
Teal is the weights (2024-09-15), amber triangles are 5 dated articles (each links to it), grey ticks are repo commits.
308 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Everything else, folded: the full tables, every source.
Qwen2.5-14B is a 14.8B parameter causal language model designed for text generation.[1][2][3] The model architecture uses transformers with RoPE, SwiGLU, RMSNorm, and Attention QKV bias.[4] It is released under the Apache 2.0 license and supports a context length of 131,072 tokens.[5][6]
Generated 2026-10-07 15:05 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
Listed in LLMs
| Pipeline tag | text-generation |
|---|---|
| Architectures | Qwen2ForCausalLM |
| Model type | qwen2 |
| Dtype | bfloat16 |
| Params by dtype | BF16: 14,770,033,664 |
| Size band | L |
| MoE | no |
| Context length | 131,072 (config.json max_position_embeddings) |
| Languages | en |
| Papers | arXiv:2407.10671 |
| License | Apache 2.0 (apache-2.0) |
| Hub snapshot | 2026-10-25 |
vLLM: supported · Qwen2ForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Qwen2ForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Weights-only memory for Qwen2.5-14B: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 35 GB · Single high-end GPU (<80 GB) |
|---|---|
| FP8 (1 byte/param) | 18 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 8.9 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
finetune: 121quantized: 78merge: 80adapter: 29
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| mlx-community/Qwen2.5-14B-Instruct-4bit | finetune | -4bit | 14.8B | L | apache-2.0 | mlx-community | 14 | 1.6M |
| unsloth/Qwen2.5-14B-bnb-4bit | quantized | -bnb | 15.2B | L | apache-2.0 | unsloth | 5 | 981.6k |
| arcee-ai/Virtuoso-Small-v2 | finetune | — | 14.8B | L | apache-2.0 | arcee-ai | 35 | 65.2k |
| arcee-ai/SuperNova-Medius | finetune | — | 14.8B | L | apache-2.0 | arcee-ai | 219 | 37.2k |
| unsloth/Qwen2.5-14B-unsloth-bnb-4bit | quantized | -bnb | 15.1B | L | apache-2.0 | unsloth | 1 | 46.9k |
| Mungert/Qwen2.5-14B-Instruct-1M-GGUF | quantized | GGUF | — | — | apache-2.0 | Mungert | 6 | 46.2k |
| unsloth/Qwen2.5-14B | finetune | — | 14.8B | L | apache-2.0 | unsloth | 2 | 42.4k |
| deepcogito/cogito-v1-preview-qwen-14B | finetune | — | 14.8B | L | apache-2.0 | deepcogito | 87 | 36.9k |
| nvidia/OpenReasoning-Nemotron-14B | finetune | — | 14.8B | L | cc-by-4.0 | nvidia | 45 | 32k |
| arcee-ai/Virtuoso-Small | finetune | — | 14.8B | L | apache-2.0 | arcee-ai | 72 | 30.6k |
| legmlai/legml-v0.1 | finetune | — | 14.8B | L | apache-2.0 | legmlai | 14 | 31.4k |
| ThomasBaruzier/Qwen2.5-14B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | ThomasBaruzier | 1 | 30.7k |
| mediainbox/cogito-14b-gptq-q4 | quantized | gptq | 14.8B | L | apache-2.0 | mediainbox | 1 | 27.8k |
| QuantFactory/medius-erebus-magnum-14b-GGUF | quantized | GGUF | — | — | — | QuantFactory | 3 | 24.4k |
| arcee-ai/Virtuoso-Small-GGUF | quantized | GGUF | — | — | apache-2.0 | arcee-ai | 11 | 23.9k |
| mradermacher/Qwen2.5-14B-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 1 | 22.3k |
| DatarusAI/Datarus-R1-14B-preview | finetune | — | 14.8B | L | apache-2.0 | DatarusAI | 141 | 14.7k |
| mradermacher/Qwen2.5-14B-i1-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 2 | 13.1k |
| nvidia/OpenMath-Nemotron-14B | finetune | — | 14.8B | L | cc-by-4.0 | nvidia | 20 | 11k |
| QuantFactory/Qwen2.5-14B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | QuantFactory | 4 | 10.9k |
| Mungert/Qwen2.5-14B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | Mungert | 6 | 10.2k |
| Mungert/OpenMath-Nemotron-14B-Kaggle-GGUF | quantized | GGUF | — | — | cc-by-4.0 | Mungert | 0 | 10.2k |
| inclusionAI/ASearcher-Local-14B | finetune | — | — | — | apache-2.0 | inclusionAI | 29 | 8.2k |
| EVA-UNIT-01/EVA-Qwen2.5-14B-v0.2 | finetune | — | 14.8B | L | apache-2.0 | EVA-UNIT-01 | 32 | 7.9k |
| nvidia/OpenMath-Nemotron-14B-Kaggle | finetune | — | 14.8B | L | cc-by-4.0 | nvidia | 21 | 8.2k |
| QuantFactory/Qwen2.5-Coder-14B-GGUF | quantized | GGUF | — | — | apache-2.0 | QuantFactory | 4 | 7.4k |
| CalamitousFelicitousness/Qwen2.5-14B-Instruct-fp8-dynamic | quantized | -fp8 | 14.8B | L | apache-2.0 | CalamitousFelicitousness | 1 | 6.7k |
| netease-youdao/Confucius4-T3PO | finetune | — | 14.8B | L | apache-2.0 | netease-youdao | 79 | 1.9k |
| VAGOsolutions/SauerkrautLM-v2-14b-SFT | finetune | — | 14.8B | L | apache-2.0 | VAGOsolutions | 9 | 5.4k |
| TIGER-Lab/General-Reasoner-Qwen2.5-14B | finetune | — | 14.8B | L | apache-2.0 | TIGER-Lab | 5 | 5.2k |
| QuantFactory/QwenTies-14B-GGUF | merge | GGUF | — | — | — | QuantFactory | 1 | 4.9k |
| QuantFactory/Qwen2.5-14B-Instruct-1M-GGUF | quantized | GGUF | — | — | apache-2.0 | QuantFactory | 3 | 4.7k |
| YOYO-AI/ZYH-LLM-Qwen2.5-14B-V4 | merge | — | 14.8B | L | apache-2.0 | YOYO-AI | 8 | 4.3k |
| Tsunami-th/Tsunami-1.0-14B-Instruct | finetune | — | 14.8B | L | apache-2.0 | Tsunami-th | 0 | 4.6k |
| wanlige/li-14b-v0.4 | merge | — | 14.8B | L | apache-2.0 | wanlige | 21 | 3.5k |
| QuantFactory/Virtuoso-Small-GGUF | quantized | GGUF | — | — | apache-2.0 | QuantFactory | 2 | 4.1k |
| QuantFactory/li-14b-v0.4-GGUF | merge | GGUF | — | — | — | QuantFactory | 1 | 3.9k |
| RedHatAI/Qwen2.5-14B-FP8-dynamic | quantized | -FP8 | 14.8B | L | apache-2.0 | RedHatAI | 2 | 3.8k |
| EVA-UNIT-01/EVA-Qwen2.5-14B-v0.1 | finetune | — | 14.8B | L | apache-2.0 | EVA-UNIT-01 | 18 | 2.7k |
| QuantFactory/14B-Qwen2.5-Freya-x1-GGUF | quantized | GGUF | — | — | other | QuantFactory | 1 | 3.4k |
| Date | Commit | SHA |
|---|---|---|
| 2024-09-20 | Upload ./tokenizer_config.json with huggingface_hub | 97e1e76 |
| 2024-09-19 | Upload ./tokenizer_config.json with huggingface_hub | d02b64b |
| 2024-09-18 | update README & LICENSE | 83a1904 |
| 2024-09-18 | Create LICENSE | 707e3fe |
| 2024-09-18 | Update README.md | 272acf2 |
| 2024-09-17 | Update README.md | ce16d88 |
| 2024-09-17 | Create README.md | 5c8d2a3 |
| 2024-09-16 | Upload model.safetensors.index.json with huggingface_hub | b00d1d8 |
| 2024-09-16 | Upload tokenizer_config.json with huggingface_hub | 8d5b6bd |
| 2024-09-15 | Upload folder using huggingface_hub | 05567cc |
| 2024-09-15 | initial commit | 5ff9380 |
Hosted endpoints observed for Qwen2.5-14B, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| Featherless hf | Qwen/Qwen2.5-14B | — | — | — | — | — | — | — | live | 2026-10-25 |
Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.
arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:37 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:05 UTC · Last verified 2026-10-25
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.
