Up to date
Qwen Apache 2.0 out 2 years ago
494M parameters, text in, text out, 32K tokens of context. 1 provider serves it.
Verified 2026-10-25 Rank #62 in LLMs · score 105.2 method v1 Sources 3 Downloads 30d 5.3M · #29 of 619 tracked
Changed this week: 19 new quantized builds on the Hub (352 in all).2026-10-24 · 18 new quantized builds on the Hub (351 in all).2026-10-24
Picked by rule from 31 tracked pieces: the lab's own post, the most shared, a field report, the latest.
The 1 pick above, plus these 24. 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 (352 quantized builds); a big base for builders (2,049 fine-tunes, adapters and merges).
Top Spaces running it: pliny-the-prompter/obliteratus ♥ 621 · allenai/reward-bench ♥ 438 · embedl/hfviewer ♥ 181 · kr-cen/Qwen-Image-MICo ♥ 76
Weights only, an estimate: leave room for the KV cache. Method
About 0.3 GB at INT4, 0.6 GB at FP8, 1.2 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 (2024-09-16), amber triangles are 5 dated articles (each links to it), grey ticks are repo commits.
2,401 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.
Qwen2.5-0.5B-Instruct is a 494M parameter instruction-tuned causal language model designed for text generation.[1][2] It supports a context length of 32,768 tokens and can generate up to 8,192 tokens.[3]
Generated 2026-10-07 15:00 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 |
|---|---|
| Library | transformers |
| Architectures | Qwen2ForCausalLM |
| Model type | qwen2 |
| Dtype | bfloat16 |
| Params by dtype | BF16: 494,032,768 |
| Size band | S |
| MoE | no |
| Context length | 32,768 (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 Qwen2.5-0.5B-Instruct: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 1.2 GB · Edge (<4 GB) |
|---|---|
| FP8 (1 byte/param) | 0.6 GB · Edge (<4 GB) |
| INT4 (0.5 bytes/param) | 0.3 GB · Edge (<4 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
Base model: Qwen/Qwen2.5-0.5B
quantized: 352finetune: 1,101merge: 44adapter: 904
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| livekit/turn-detector | quantized | — | 135M | S | other | livekit | 166 | 6.4M |
| unsloth/Qwen2.5-0.5B-Instruct-bnb-4bit | quantized | -bnb | 505M | S | apache-2.0 | unsloth | 4 | 2.4M |
| Qwen/Qwen2.5-0.5B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | Qwen | 157 | 1.9M |
| mlc-ai/Qwen2.5-0.5B-Instruct-q4f16_1-MLC | quantized | — | — | — | — | mlc-ai | 4 | 1.7M |
| unsloth/Qwen2.5-0.5B-Instruct | finetune | — | 494M | S | apache-2.0 | unsloth | 14 | 1.5M |
| unsloth/Qwen2.5-0.5B-Instruct-unsloth-bnb-4bit | quantized | -bnb | 504M | S | apache-2.0 | unsloth | 6 | 716.1k |
| OpenGVLab/InternVL2_5-1B | merge | — | 938M | S | mit | OpenGVLab | 71 | 564.2k |
| bartowski/Qwen2.5-0.5B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | bartowski | 19 | 447.6k |
| OPEA/Qwen2.5-0.5B-Instruct-int4-sym-inc | quantized | int4 | 184M | S | apache-2.0 | OPEA | 0 | 188.8k |
| onnx-community/Qwen2.5-0.5B-Instruct | quantized | — | — | — | — | onnx-community | 20 | 186.3k |
| Qwen/Qwen2.5-0.5B-Instruct-AWQ | quantized | AWQ | 630M | S | apache-2.0 | Qwen | 17 | 159.7k |
| lmstudio-community/Qwen2.5-0.5B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | lmstudio-community | 8 | 139k |
| manycore-research/SpatialLM-Qwen-0.5B | finetune | — | 505M | S | cc-by-nc-4.0 | manycore-research | 94 | 121.5k |
| taobao-mnn/Qwen2.5-0.5B-Instruct-MNN | finetune | — | — | — | apache-2.0 | taobao-mnn | 4 | 121.9k |
| lmstudio-community/Qwen2.5-0.5B-Instruct-MLX-4bit | quantized | MLX | 494M | S | apache-2.0 | lmstudio-community | 1 | 115.1k |
| Qwen/Qwen2.5-0.5B-Instruct-GPTQ-Int4 | quantized | GPTQ | 494M | S | apache-2.0 | Qwen | 15 | 114.3k |
| manycore-research/SpatialLM1.1-Qwen-0.5B | finetune | — | 604M | S | cc-by-nc-4.0 | manycore-research | 38 | 105.7k |
| litert-community/Qwen2.5-0.5B-Instruct | finetune | — | — | — | apache-2.0 | litert-community | 9 | 77.9k |
| Qwen/Qwen2.5-0.5B-Instruct-GPTQ-Int8 | quantized | GPTQ | 494M | S | apache-2.0 | Qwen | 17 | 62.1k |
| ReDiX/Qwen2.5-0.5B-Instruct-ITA | finetune | — | — | — | apache-2.0 | ReDiX | 3 | 47.1k |
| madhurjindal/Jailbreak-Detector-2-XL | adapter | — | — | — | mit | madhurjindal | 17 | 42.3k |
| Vikhrmodels/Vikhr-Qwen-2.5-0.5b-Instruct | finetune | — | 494M | S | apache-2.0 | Vikhrmodels | 19 | 34.3k |
| lightblue/lb-reranker-0.5B-v1.0 | finetune | — | 494M | S | apache-2.0 | lightblue | 76 | 30.3k |
| nvidia/Eagle2-1B | merge | — | 1.06B | S | cc-by-nc-4.0 | nvidia | 33 | 27.8k |
| legraphista/Qwen2.5-0.5B-Instruct-IMat-GGUF | quantized | GGUF | — | — | apache-2.0 | legraphista | 0 | 27.1k |
| IAAR-Shanghai/xVerify-0.5B-I | finetune | — | 494M | S | cc-by-nc-nd-4.0 | IAAR-Shanghai | 11 | 21.8k |
| gaianet/Qwen2.5-0.5B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | gaianet | 0 | 21.8k |
| mlc-ai/Qwen2.5-0.5B-Instruct-q4f32_1-MLC | quantized | — | — | — | — | mlc-ai | 0 | 21.5k |
| KingNish/Qwen2.5-0.5b-Test-ft | finetune | — | 494M | S | apache-2.0 | KingNish | 12 | 20k |
| wuyanzu4692/task-13-Qwen-Qwen2.5-0.5B-Instruct | adapter | — | — | — | — | wuyanzu4692 | 0 | 19.4k |
| Ihor/Text2Graph-R1-Qwen2.5-0.5b | finetune | — | 494M | S | apache-2.0 | Ihor | 24 | 17.9k |
| second-state/Qwen2.5-0.5B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | second-state | 2 | 17.4k |
| mradermacher/Qwen2.5-0.5B-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 0 | 15.6k |
| yiyuliu/voice-command-intent-qwen0.5b | adapter | GGUF | — | — | apache-2.0 | yiyuliu | 0 | 14.4k |
| mradermacher/Qwen2.5-0.5B-Instruct-i1-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 2 | 14.1k |
| furiosa-ai/Qwen2.5-0.5B-Instruct | finetune | — | — | — | apache-2.0 | furiosa-ai | 0 | 13.3k |
| Hugggme/Qwen2.5-0.5B-Instruct-Q4_K_M-GGUF | quantized | GGUF | — | — | apache-2.0 | Hugggme | 26 | 10.5k |
| sallani/EUAIAct-Qwen2.5-0.5B-Edge | adapter | GGUF | 494M | S | apache-2.0 | sallani | 2 | 11.5k |
| reaperdoesntknow/DeepReasoning_1R | finetune | — | 494M | S | — | reaperdoesntknow | 2 | 11.4k |
| saidutta69/Qwen2.5-0.5B-Instruct-heretic | quantized | GGUF | 494M | S | other | saidutta69 | 0 | 11k |
| Date | Commit | SHA |
|---|---|---|
| 2024-09-25 | Update metadata: Add library_name to Transformers (#2) | 7ae5576 |
| 2024-09-19 | Upload ./tokenizer_config.json with huggingface_hub | a8b602d |
| 2024-09-18 | update README & LICENSE | f0ef055 |
| 2024-09-18 | Update tokenizer_config.json | a338b55 |
| 2024-09-17 | Update README.md | d955144 |
| 2024-09-17 | Update README.md | be08fe9 |
| 2024-09-17 | Update README.md | 361e0b3 |
| 2024-09-17 | Create README.md | a0f8e5d |
| 2024-09-16 | Upload folder using huggingface_hub | eaa56b5 |
| 2024-09-16 | initial commit | 4a7e54c |
Hosted endpoints observed for Qwen2.5-0.5B-Instruct, 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/Qwen2.5-0.5B-Instruct | — | — | — | — | — | — | — | 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 21:33 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:00 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.
