Up to date
deepseek-ai MIT out 21 months ago
7.62B parameters, text in, text out, 128K tokens of context. 2 providers serve it from $0.15 per million input tokens.
Verified 2026-10-26 07:27 UTC Rank #88 in LLMs · score 101.2 method v1 Sources 3 Downloads 30d 309.1k · #216 of 619 tracked
Changed this week: 3 new quantized builds on the Hub (204 in all).2026-10-24 · 1 new provider: Nscale.2026-10-23
Picked by rule from 34 tracked pieces: the lab's own post, the most shared, a field report, the latest.
Zyphra released ZAYA1-8B, a model that matches DeepSeek-R1 on math benchmarks while using less than 1 billion active parameters.
The 2 picks above, plus these 23. 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.
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 2 API providers; heavily run locally (203 quantized builds); a big base for builders (540 fine-tunes, adapters and merges).
Top Spaces running it: pliny-the-prompter/obliteratus ♥ 621 · ginigen-ai/smol-worldcup ♥ 59 · KBaba7/Quant ♥ 51 · ruslanmv/DeepSeek-R1-Chatbot ♥ 36
Weights only, an estimate: leave room for the KV cache. Method
About 4.6 GB at INT4, 9.1 GB at FP8, 18 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 (2025-01-20), amber triangles are 10 dated articles (each links to it), grey ticks are repo commits.
743 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.
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported AIME 2024 · Reason through hard problems | 55.5 pass@1 | DeepSeek-R1-Distill-Qwen-7B | model author | 2026-10-07 |
| author-reported AIME 2024 · Reason through hard problems | 83.3 other | DeepSeek-R1-Distill-Qwen-7B | model author | 2026-10-07 |
| author-reported Codeforces · Write code with agents | 1189 rating | DeepSeek-R1-Distill-Qwen-7B | model author | 2026-10-07 |
| author-reported GPQA Diamond · Reason through hard problems | 49.1 pass@1 | DeepSeek-R1-Distill-Qwen-7B | model author | 2026-10-07 |
| author-reported LiveCodeBench · Write code with agents | 37.6 pass@1 | DeepSeek-R1-Distill-Qwen-7B | model author | 2026-10-07 |
| author-reported MATH-500 · Reason through hard problems | 92.8 pass@1 | DeepSeek-R1-Distill-Qwen-7B | model author | 2026-10-07 |
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: 7,615,616,512 |
| Size band | M |
| MoE | no |
| Context length | 131,072 (config.json max_position_embeddings) |
| Papers | arXiv:2501.12948 |
| License | MIT (mit) |
| Hub snapshot | 2026-10-26 |
vLLM: supported · Qwen2ForCausalLM · 4b1c04e76fe9ac3ca8b7e541dd423e7d5f38dc60SGLang: supported · Qwen2ForCausalLM · 60d941293e3af27a12c6eb64e2c3e2e1a17f100f
vLLM registry (4b1c04e76fe9ac3ca8b7e541dd423e7d5f38dc60) · SGLang registry (60d941293e3af27a12c6eb64e2c3e2e1a17f100f) · 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 DeepSeek-R1-Distill-Qwen-7B: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 18 GB · Single GPU (<24 GB) |
|---|---|
| FP8 (1 byte/param) | 9.1 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 4.6 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
quantized: 203finetune: 314merge: 64adapter: 162
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| lmstudio-community/DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | — | lmstudio-community | 90 | 1.8M |
| bartowski/DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | — | bartowski | 149 | 1.5M |
| unsloth/DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | apache-2.0 | unsloth | 111 | 971.8k |
| mlx-community/DeepSeek-R1-Distill-Qwen-7B-8bit | quantized | -8bit | 7.62B | M | — | mlx-community | 8 | 592.7k |
| unsloth/DeepSeek-R1-Distill-Qwen-7B-unsloth-bnb-4bit | quantized | -bnb | 7.76B | M | apache-2.0 | unsloth | 28 | 414.8k |
| unsloth/DeepSeek-R1-Distill-Qwen-7B | finetune | — | 7.62B | M | apache-2.0 | unsloth | 20 | 96.2k |
| mlx-community/DeepSeek-R1-Distill-Qwen-7B-4bit | quantized | -4bit | 7.62B | M | — | mlx-community | 25 | 78.6k |
| unsloth/DeepSeek-R1-Distill-Qwen-7B-bnb-4bit | quantized | -bnb | 7.82B | M | apache-2.0 | unsloth | 3 | 75.1k |
| mlx-community/DeepSeek-R1-Distill-Qwen-7B-6bit | quantized | — | 7.62B | M | — | mlx-community | 2 | 60k |
| RedHatAI/DeepSeek-R1-Distill-Qwen-7B-quantized.w8a8 | quantized | -quant | 7.62B | M | mit | RedHatAI | 5 | 52.2k |
| mlc-ai/DeepSeek-R1-Distill-Qwen-7B-q4f16_1-MLC | quantized | — | — | — | — | mlc-ai | 2 | 44.2k |
| mradermacher/DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | mit | mradermacher | 1 | 41.1k |
| GenPRM/GenPRM-7B | finetune | — | 7.62B | M | mit | GenPRM | 6 | 38.7k |
| RedHatAI/DeepSeek-R1-Distill-Qwen-7B-FP8-dynamic | quantized | -FP8 | 7.62B | M | mit | RedHatAI | 1 | 32.5k |
| DevQuasar-7/deepseek-ai.DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | — | DevQuasar-7 | 2 | 31.8k |
| nvidia/DLER-R1-7B-Research | finetune | — | 7.62B | M | — | nvidia | 17 | 26.5k |
| eaddario/DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | mit | eaddario | 3 | 26.4k |
| ngxson/DeepSeek-R1-Distill-Qwen-7B-abliterated-GGUF | quantized | GGUF | — | — | — | ngxson | 12 | 25.5k |
| Skywork/Skywork-OR1-Math-7B | finetune | — | 7.62B | M | — | Skywork | 19 | 24.8k |
| Skywork/Skywork-OR1-7B | finetune | — | 7.62B | M | — | Skywork | 14 | 20k |
| jakiAJK/DeepSeek-R1-Distill-Qwen-7B_AWQ | quantized | AWQ | 7.62B | M | — | jakiAJK | 2 | 19.6k |
| OpenVINO/DeepSeek-R1-Distill-Qwen-7B-int4-ov | quantized | int4 | — | — | mit | OpenVINO | 2 | 16.3k |
| lightblue/DeepSeek-R1-Distill-Qwen-7B-Japanese | finetune | — | 7.62B | M | apache-2.0 | lightblue | 33 | 14.6k |
| roleplaiapp/DeepSeek-R1-Distill-Qwen-7B-Q4_K_M-GGUF | quantized | GGUF | — | — | — | roleplaiapp | 3 | 15.3k |
| ModelCloud/DeepSeek-R1-Distill-Qwen-7B-gptqmodel-4bit-vortex-v2 | quantized | gptq | 7.62B | M | apache-2.0 | ModelCloud | 8 | 14.9k |
| huihui-ai/DeepSeek-R1-Distill-Qwen-7B-abliterated-v2 | finetune | — | 7.62B | M | — | huihui-ai | 56 | 12.2k |
| jakiAJK/DeepSeek-R1-Distill-Qwen-7B_GPTQ-int4 | quantized | GPTQ | 7.62B | M | — | jakiAJK | 3 | 14.6k |
| RedHatAI/DeepSeek-R1-Distill-Qwen-7B-quantized.w4a16 | quantized | -quant | 7.62B | M | mit | RedHatAI | 2 | 14.3k |
| second-state/DeepSeek-R1-Distill-Qwen-7B-GGUF | quantized | GGUF | — | — | apache-2.0 | second-state | 1 | 13.6k |
| Kondara/DeepSeek-R1-Distill-Qwen-7B-Q4_K_M-GGUF | quantized | GGUF | — | — | mit | Kondara | 4 | 12.7k |
| ThomasBaruzier/DeepSeek-R1-ReDistill-Qwen-7B-v1.1-GGUF | quantized | GGUF | — | — | mit | ThomasBaruzier | 0 | 12.3k |
| qihoo360/Light-R1-7B-DS | finetune | — | 7.62B | M | apache-2.0 | qihoo360 | 22 | 10.8k |
| mlx-community/DeepSeek-R1-Distill-Qwen-7B | finetune | — | 7.62B | M | — | mlx-community | 6 | 11.5k |
| arcee-ai/Arcee-Maestro-7B-Preview | finetune | — | 7.61B | M | apache-2.0 | arcee-ai | 38 | 9.8k |
| Andycurrent/DeepSeek-R1-Distill-Qwen-7B-Uncensored_GGUF | quantized | GGUF | — | — | apache-2.0 | Andycurrent | 6 | 11.2k |
| mradermacher/DeepSeek-R1-Distill-Qwen-7B-i1-GGUF | quantized | GGUF | — | — | mit | mradermacher | 2 | 11.3k |
| saidutta69/DeepSeek-R1-Distill-Qwen-7B-heretic | quantized | GGUF | 7.62B | M | other | saidutta69 | 1 | 11.1k |
| mmnga/DeepSeek-R1-Distill-Qwen-7B-gguf | quantized | gguf | — | — | mit | mmnga | 3 | 10.1k |
| sabafallah/DeepSeek-R1-Distill-Qwen-7B-Q4_K_M-GGUF | quantized | GGUF | — | — | mit | sabafallah | 0 | 9.9k |
| mlx-community/DeepSeek-R1-Distill-Qwen-7B-3bit | quantized | — | 7.62B | M | — | mlx-community | 0 | 9.8k |
| Date | Commit | SHA |
|---|---|---|
| 2025-02-24 | Update bib info | 916b56a |
| 2025-02-09 | Update tokenizer_config.json | 6602cad |
| 2025-02-01 | Update README.md | 14dd113 |
| 2025-01-26 | Update README.md | 09db661 |
| 2025-01-23 | Update README.md | 393119f |
| 2025-01-22 | Update tokenizer_config.json | 4831ee1 |
| 2025-01-22 | Update tokenizer_config.json | b640c47 |
| 2025-01-21 | Add generation_config.json | 4e5485e |
| 2025-01-21 | add library tag for better code snippets and tags (#1) | 10f2fc8 |
| 2025-01-20 | Release DeepSeek-R1 | 008b8c2 |
| 2025-01-20 | Add files using upload-large-folder tool | a7468ad |
| 2025-01-20 | initial commit | 0802d82 |
Hosted endpoints observed for DeepSeek-R1-Distill-Qwen-7B, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.15/M.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| Nscale hf | deepseek-ai/DeepSeek-R1-Distill-Qwen-7B | 128K | — | $0.15 | $0.15 | 179 | 570 ms | no | live | 2026-10-26 |
| Featherless hf | deepseek-ai/DeepSeek-R1-Distill-Qwen-7B | — | — | — | — | — | — | — | live | 2026-10-26 |
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: 2026-10-26 · HF API: 2026-10-26 07:15 UTC · Providers: 2026-10-26 07:26 UTC · Articles: 2026-10-26 07:27 UTC · Summary: 2026-10-07 15:01 UTC · Last verified 2026-10-26 07:27 UTC
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.
