Up to date
Qwen Apache 2.0 out 18 months ago
8.19B parameters, text in, text out, 32K tokens of context. 1 provider serves it.
Verified 2026-10-25 Rank #185 in LLMs · score 89.8 method v1 Sources 3 Downloads 30d 733.2k · #136 of 619 tracked
Changed this week: 1 new quantized build on the Hub (51 in all).2026-10-24
Picked by rule from 18 tracked pieces: the lab's own post, the most shared, a field report, the latest.
OpenAI released gpt-oss-120b and gpt-oss-20b, their first open-weight models since GPT-2 in 2019.
The 4 picks above, plus these 14. 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 (51 quantized builds); a big base for builders (798 fine-tunes, adapters and merges).
Top Spaces running it: dqy08/InfoLens1 ♥ 22 · openeurollm/prelude-evaluations ♥ 10 · Lunamos/flas-demo ♥ 2 · deniskondratev/autogenesis-chat ♥ 1
Weights only, an estimate: leave room for the KV cache. Method
About 4.9 GB at INT4, 9.8 GB at FP8, 20 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 (2025-04-28), amber triangles are 8 dated articles (each links to it), grey ticks are repo commits.
849 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 | Qwen3ForCausalLM |
| Model type | qwen3 |
| Dtype | bfloat16 |
| Params by dtype | BF16: 8,190,735,360 |
| Size band | M |
| MoE | no |
| Context length | 32,768 (config.json max_position_embeddings) |
| Papers | arXiv:2505.09388 |
| License | Apache 2.0 (apache-2.0) |
| Hub snapshot | 2026-10-25 |
vLLM: supported · Qwen3ForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Qwen3ForCausalLM · 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 Qwen3-8B-Base: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 20 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 9.8 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 4.9 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
quantized: 51finetune: 636adapter: 148merge: 14
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| Qwen/Qwen3-Embedding-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | Qwen | 155 | 369.3k |
| unsloth/Qwen3-8B-Base | finetune | — | 8.19B | M | apache-2.0 | unsloth | 5 | 130.6k |
| unsloth/Qwen3-8B-Base-unsloth-bnb-4bit | quantized | -bnb | 8.39B | M | apache-2.0 | unsloth | 4 | 107.4k |
| unsloth/Qwen3-8B-Base-bnb-4bit | quantized | -bnb | 8.41B | M | apache-2.0 | unsloth | 3 | 46.5k |
| tomaarsen/Qwen3-Reranker-8B-seq-cls | finetune | — | 7.57B | M | apache-2.0 | tomaarsen | 11 | 38.3k |
| Mungert/Qwen3-Embedding-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | Mungert | 4 | 18.2k |
| hkust-nlp/drkernel-8b-coldstart | finetune | — | 256M | S | — | hkust-nlp | 0 | 14.4k |
| unsloth/Qwen3-Embedding-8B | finetune | — | 7.57B | M | apache-2.0 | unsloth | 7 | 12.8k |
| Mungert/Qwen3-Reranker-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | Mungert | 4 | 12k |
| mlx-community/Qwen3-Embedding-8B-mxfp8 | finetune | — | — | — | apache-2.0 | mlx-community | 4 | 11.3k |
| cmu-lti/osim-8b-mid | finetune | — | 8.19B | M | mit | cmu-lti | 2 | 11k |
| Alphatao/Affine-9711767 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 11k |
| snuh/hari-q3-8b | finetune | — | 8.19B | M | apache-2.0 | snuh | 6 | 10.3k |
| QuantFactory/Qwen3-Reranker-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | QuantFactory | 4 | 10.1k |
| Alphatao/Affine-5956831 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 10.2k |
| hkust-nlp/drkernel-8b | finetune | — | 8.19B | M | — | hkust-nlp | 5 | 9.8k |
| AlphaGaO/Qwen3-8B-GPTQ | quantized | GPTQ | 8.19B | M | apache-2.0 | AlphaGaO | 1 | 9.7k |
| Alphatao/Affine-1855255 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 9.6k |
| McGill-NLP/AfriqueQwen-8B | finetune | — | 8.19B | M | cc-by-4.0 | McGill-NLP | 2 | 9.2k |
| Mungert/Qwen3-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | Mungert | 7 | 8.4k |
| PrimeIntellect/Qwen3-8B | finetune | — | 8.19B | M | apache-2.0 | PrimeIntellect | 0 | 8.5k |
| Alphatao/Affine-9459823 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 8.3k |
| Alphatao/Affine-1901852 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 8.1k |
| mradermacher/Qwen3-8B-Base-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 3 | 7.5k |
| Alphatao/Affine-2501551 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 7.2k |
| zeromodels/qwen3-8b-base | finetune | — | — | — | apache-2.0 | zeromodels | 0 | 7k |
| codys12/Qwen3-8B-BitNet | quantized | — | 2.52B | M | apache-2.0 | codys12 | 17 | 5.9k |
| mradermacher/Qwen3-8B-Base-i1-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 1 | 6.6k |
| nate-rahn/0713-sft_category_generation-qwen3_8b_base | finetune | — | 1.02B | S | apache-2.0 | nate-rahn | 0 | 6.3k |
| Alphatao/Affine-2333827 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 6k |
| OpenDataArena/Qwen3-8B-ODA-Math-460k | finetune | — | 308K | S | cc-by-nc-4.0 | OpenDataArena | 2 | 5.4k |
| QuantFactory/Qwen3-8B-GGUF | quantized | GGUF | — | — | apache-2.0 | QuantFactory | 2 | 5k |
| Alphatao/Affine-7470548 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 4.5k |
| Alphatao/Affine-5246433 | finetune | — | 8.19B | M | apache-2.0 | Alphatao | 0 | 4.4k |
| stepfun-ai/RLVR-8B-0926 | finetune | — | 8.19B | M | mit | stepfun-ai | 9 | 3.9k |
| tencent/POINTS-GUI-G | finetune | — | 8.91B | M | other | tencent | 22 | 3.2k |
| Antigma/Qwen3-8B-Base-GGUF | quantized | GGUF | — | — | apache-2.0 | Antigma | 0 | 4.1k |
| HealthDataHub/PARTAGES-Qwen3-8B-PDAPT-SLERP | finetune | — | 8.19B | M | apache-2.0 | HealthDataHub | 0 | 3.8k |
| p-e-w/Qwen3-8B-heretic | finetune | — | 8.19B | M | apache-2.0 | p-e-w | 4 | 3.4k |
| Qwen/SAE-Res-Qwen3-8B-Base-W64K-L0_50 | finetune | — | — | — | other | Qwen | 14 | 2.8k |
| Date | Commit | SHA |
|---|---|---|
| 2025-05-21 | Update README.md | 49e3418 |
| 2025-05-20 | update tokenizer_config.json | b86f722 |
| 2025-04-28 | Update README.md | b53affe |
| 2025-04-28 | Update README.md | 11de6bf |
| 2025-04-28 | Update tokenizer_config.json | a64df03 |
| 2025-04-28 | Upload merges.txt | ac2b9a2 |
| 2025-04-28 | Update README.md | 753f1e2 |
| 2025-04-28 | Upload folder using huggingface_hub | dfae426 |
| 2025-04-28 | initial commit | ce6d71a |
Hosted endpoints observed for Qwen3-8B-Base, 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/Qwen3-8B-Base | — | — | — | — | — | — | — | 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 21:36 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:03 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.
