Up to date
Qwen Research / non-commercial (tongyi-qianwen-research) out 3 years ago
620M parameters, text in, text out, 32K tokens of context. 1 provider serves it.
Verified 2026-10-25 Rank #316 in LLMs · score 82.7 method v1 Sources 3 Downloads 30d 56.9k · #372 of 619 tracked
Picked by rule from 7 tracked pieces: the lab's own post, the most shared, a field report, the latest.
Qwen1.5-MoE-A2.7B matches 7B model performance with only 2.7 billion activated parameters using a 64-expert configuration.
The 2 picks above, plus these 5. 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; 37 quantized builds for local use; a big base for builders (301 fine-tunes, adapters and merges).
Top Spaces running it: lhoestq/fake-data-generator-jsonl ♥ 68 · gsarti/pecore ♥ 18 · seanpedrickcase/Light-PDF-Web-QA-Chatbot ♥ 14 · Justinrune/LLaMA-Factory ♥ 12
Weights only, an estimate: leave room for the KV cache. Method
About 0.4 GB at INT4, 0.7 GB at FP8, 1.5 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 (2024-01-31), amber triangles are 3 dated articles (each links to it), grey ticks are repo commits.
338 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.
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 |
|---|---|
| Library | transformers |
| Architectures | Qwen2ForCausalLM |
| Model type | qwen2 |
| Dtype | bfloat16 |
| Params by dtype | BF16: 619,570,176 |
| Size band | S |
| MoE | no |
| Context length | 32,768 (config.json max_position_embeddings) |
| Languages | en |
| Papers | arXiv:2309.16609 |
| License | Research / non-commercial (tongyi-qianwen-research) (other) |
| 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 Qwen1.5-0.5B-Chat: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 1.5 GB · Edge (<4 GB) |
|---|---|
| FP8 (1 byte/param) | 0.7 GB · Edge (<4 GB) |
| INT4 (0.5 bytes/param) | 0.4 GB · Edge (<4 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
quantized: 37finetune: 19adapter: 282
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| Xenova/Qwen1.5-0.5B-Chat | quantized | — | — | — | — | Xenova | 9 | 51.1k |
| second-state/Qwen1.5-0.5B-Chat-GGUF | quantized | GGUF | — | — | other | second-state | 3 | 16.6k |
| mlc-ai/Qwen1.5-0.5B-Chat-q4f16_1-MLC | quantized | — | — | — | — | mlc-ai | 0 | 8.9k |
| Felladrin/gguf-Qwen1.5-0.5B-Chat | quantized | gguf | — | — | — | Felladrin | 2 | 7.6k |
| DevQuasar-9/Qwen.Qwen1.5-0.5B-Chat-GGUF | quantized | GGUF | — | — | — | DevQuasar-9 | 1 | 5k |
| mradermacher/Qwen1.5-0.5B-Chat-i1-GGUF | quantized | GGUF | — | — | other | mradermacher | 0 | 4.9k |
| mradermacher/Qwen1.5-0.5B-Chat-GGUF | quantized | GGUF | — | — | other | mradermacher | 0 | 2.9k |
| Felladrin/gguf-sharded-Qwen1.5-0.5B-Chat | quantized | gguf | — | — | other | Felladrin | 0 | 2.5k |
| Zynerji/Ektome-Qwen1.5-0.5B-Chat-PristinelyUncensored | quantized | GGUF | 464M | S | apache-2.0 | Zynerji | 0 | 2.1k |
| arzaan789/qwen1.5-0.5b-uncensored | quantized | GGUF | — | — | apache-2.0 | arzaan789 | 1 | 1.8k |
| onnx-community/Qwen1.5-0.5B-Chat-ONNX | quantized | — | — | — | — | onnx-community | 0 | 1.2k |
| FINGU-AI/Qwen-Orpo-v1 | finetune | — | 464M | S | apache-2.0 | FINGU-AI | 0 | 978 |
| mlc-ai/Qwen1.5-0.5B-Chat-q3f16_1-MLC | quantized | — | — | — | — | mlc-ai | 0 | 801 |
| Elaine5/Qwen1.5-0.5B-Chat-Q4_K_M-GGUF | quantized | GGUF | — | — | other | Elaine5 | 0 | 736 |
| mlc-ai/Qwen1.5-0.5B-Chat-q0f16-MLC | quantized | — | — | — | — | mlc-ai | 0 | 406 |
| itlwas/Qwen1.5-0.5B-Chat-Q4_K_M-GGUF | quantized | GGUF | — | — | other | itlwas | 0 | 305 |
| markab/Qwen1.5-Capybara-0.5B-Chat | adapter | — | 464M | S | other | markab | 1 | 226 |
| amai-gsu/Qwen1.5-0.5B-Chat-Q4_K_M-GGUF | quantized | GGUF | — | — | other | amai-gsu | 0 | 275 |
| Yermalovich/results | finetune | — | 464M | S | other | Yermalovich | 0 | 262 |
| attention-avengers/Qwen1.5-0.5B-Chat-EPFL-cDPO | adapter | — | — | — | — | attention-avengers | 0 | 260 |
| panzarasa/qwen1.5-0.5b-websight | adapter | — | 464M | S | other | panzarasa | 0 | 224 |
| Super007/aversi50 | finetune | — | 464M | S | other | Super007 | 0 | 197 |
| Anthonyg5005/Qwen1.5-0.5B-Chat-exl2 | quantized | exl2 | — | — | other | Anthonyg5005 | 0 | 171 |
| thenewth/Qwen1.5-0.5B-Chat-GrammarCorrection | adapter | — | 464M | S | apache-2.0 | thenewth | 0 | 142 |
| attention-avengers/Qwen1.5-0.5B-Chat-EPFL-ORCA-cDPO | adapter | — | — | — | — | attention-avengers | 0 | 137 |
| nblinh63/d7dec48d-6963-4941-a666-82afe3a01311 | adapter | — | — | — | other | nblinh63 | 0 | 131 |
| terry69/qwen_0.5_feedback_dirty | finetune | — | 464M | S | other | terry69 | 0 | 126 |
| lescurej/Qwen1.5-0.5B-Chat-Q4_K_M-GGUF | quantized | GGUF | — | — | other | lescurej | 0 | 118 |
| vermoney/968ff187-2ae1-4d9b-9fd9-9af8f08fe27b | quantized | — | — | — | — | vermoney | 0 | 84 |
| chkim123/Qwen1.5-0.5B-Chat-ONNX | quantized | — | — | — | — | chkim123 | 0 | 84 |
| vermoney/04ba9e7e-632e-4981-a1dd-9e4604430ba1 | quantized | — | — | — | — | vermoney | 0 | 82 |
| kokovova/a5f58fdb-fe66-4332-bf8f-1066cb6463b7 | quantized | — | — | — | — | kokovova | 0 | 82 |
| dzanbek/c25b500e-4f01-4d5e-b6b8-dbf1683fcb1c | quantized | — | — | — | — | dzanbek | 0 | 78 |
| terry69/qwen_feedback_dirty_0.5 | finetune | — | 464M | S | other | terry69 | 0 | 77 |
| Org-Exp/Hyperthink-exp | finetune | — | 494M | S | — | Org-Exp | 0 | 77 |
| mimimimihugh/claim-agent-lora | adapter | — | — | — | — | mimimimihugh | 0 | 76 |
| QINchampion/output | finetune | — | 464M | S | other | QINchampion | 0 | 75 |
| emmabhl/Qwen1.5-0.5B-Chat-ORCA-cDPO | adapter | — | — | — | — | emmabhl | 0 | 71 |
| amai-gsu/Qwen1.5-0.5B-Chat-Q4_K_S-GGUF | quantized | GGUF | — | — | other | amai-gsu | 0 | 71 |
| attention-avengers/Qwen1.5-0.5B-Chat-EPFL-ORCA-DPO | adapter | — | — | — | — | attention-avengers | 0 | 70 |
| Date | Commit | SHA |
|---|---|---|
| 2024-04-30 | update tokenizer_config.json | 4d14e38 |
| 2024-04-05 | Update README.md | f82bd36 |
| 2024-04-05 | Update README.md | 23b9054 |
| 2024-04-04 | update tokenizer_config.json | 2a01ce8 |
| 2024-02-28 | update generation_config.json | 6c70598 |
| 2024-02-20 | add torch_dtype | 7f630fd |
| 2024-02-09 | update eos token | 81b0090 |
| 2024-02-06 | update tokenizer_config.json | 6114e9c |
| 2024-02-05 | update readme | 3f5adc4 |
| 2024-02-05 | update readme | 4cbf541 |
| 2024-02-05 | Update README.md | 24a5e87 |
| 2024-02-05 | update readme | 2137f0c |
| 2024-02-05 | Update README.md | d9f867e |
| 2024-02-05 | Update README.md | b6eb0ef |
| 2024-02-05 | Update README.md | e6f98dd |
| 2024-02-05 | Update README.md | 72bc460 |
| 2024-02-03 | Update config.json | 7ff50b2 |
| 2024-02-03 | Update config.json | 3c812c8 |
| 2024-02-02 | Update generation_config.json | 6c26ede |
| 2024-02-02 | Update README.md | 5c5635f |
| 2024-01-31 | Update README.md | 25e33b3 |
| 2024-01-31 | Update README.md | ab0b543 |
| 2024-01-31 | Update README.md | cffcc20 |
| 2024-01-31 | Update README.md | 9e64766 |
| 2024-01-31 | Create LICENSE | 105896a |
| 2024-01-31 | Upload folder using huggingface_hub | b79d0c9 |
| 2024-01-31 | initial commit | a059837 |
Hosted endpoints observed for Qwen1.5-0.5B-Chat, 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/Qwen1.5-0.5B-Chat | — | — | — | — | — | — | — | 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.
| Title | Source | Published | First seen | Signals |
|---|---|---|---|---|
| Qwen1.5-110B | qwenlm.github.io official | 2024-04-26 | 2026-10-07 | hn |
| Qwen1.5-110B Preview | twitter.com | 2024-04-20 | 2026-10-07 | hn |
| Qwen1.5-32B: Fitting the Capstone of the Qwen1.5 Language Model Series | qwenlm.github.io official | 2024-04-05 | 2026-10-07 | hn |
| Qwen1.5-Moe: Matching 7B Model Performance with 1/3 Activated Parameters | qwenlm.github.io official | 2024-03-28 | 2026-10-07 | hn |
| Run Alicloud's Qwen LLM with a 2MB cross GPU/CPU rust app | secondstate.io | 2024-02-28 | 2026-10-07 | hn |
| Complete tutorial: Run Qwen LLM across your local machines without Python/C++ | secondstate.io | 2024-02-22 | 2026-10-07 | hn |
| Qwen1.5 | qwenlm.github.io official | 2024-02-05 | 2026-10-07 | hn |
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 21:38 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:08 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.
