Up to date
deepseek-ai MIT out 8 days ago
763B parameters, images and text in, text out, 1M tokens of context. 5 providers serve it from $0.30 per million input tokens.
Verified 2026-10-25 Rank #44 in LLMs · score 108.8 method v1 Sources 4 Downloads 30d 73.2k · #352 of 619 tracked
Changed this week: DeepSeek V4.1 Flash 1017 entered the release radar as served, open-weights (Hugging Face, 2026-10-23).2026-10-23
Picked by rule from 35 tracked pieces: the lab's own post, the most shared, a field report, the latest.
DeepSeek V4.1 Flash 1017 (Max) is a 552B parameter Mixture of Experts model with 16B active parameters released under the MIT license.
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.
4 providers with a live price, from $0.30 to $0.30 per million input tokens; the median is $0.30. One dot per provider, cheapest in teal.
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.
Widely served (5 API providers); heavily run locally (122 quantized builds).
Weights only, an estimate: leave room for the KV cache. Method
About 458 GB at INT4, 916 GB at FP8, 1.83 TB 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 (2026-10-17), amber triangles are 10 dated articles (each links to it).
161 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 | image-text-to-text |
|---|---|
| Library | transformers |
| Architectures | DeepseekV41ForCausalLM |
| Model type | deepseek_v41 |
| Params by dtype | BF16: 1,976,441,856 · F32: 42,307,282 · F8_E4M3: 204,015,223,296 · I8: 557,171,343,360 |
| Size band | XXL |
| MoE | no |
| Context length | 1,048,576 (provider listing) |
| License | MIT (mit) |
| Hub snapshot | 2026-10-25 |
Weights-only memory for DeepSeek-V4.1-Flash-1017: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 1.83 TB · Cluster (≥640 GB) |
|---|---|
| FP8 (1 byte/param) | 916 GB · Cluster (≥640 GB) |
| INT4 (0.5 bytes/param) | 458 GB · Multi-GPU node (<640 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
quantized: 122finetune: 37merge: 1adapter: 1
Hosted endpoints observed for DeepSeek-V4.1-Flash-1017, 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: $0.20/M.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| DeepInfra hf | deepseek-ai/DeepSeek-V4.1-Flash-1017 | 1M | — | $0.20 | $0.60 | 58 | 438 ms | yes | staging | 2026-10-25 |
| Novita hf | deepseek/deepseek-v4.1-flash-1017 | 1M | — | $0.30 | $1.20 | 99.5 | 1029 ms | yes | live | 2026-10-25 |
| Together AI hf | deepseek-ai/DeepSeek-V4.1-Flash-1017 | 1M | — | $0.30 | $1.20 | 308.8 | 171 ms | yes | live | 2026-10-25 |
| Baseten hf | deepseek-ai/DeepSeek-V4.1-Flash-1017 | 1M | — | $0.30 | $1.20 | 123 | 4182 ms | yes | live | 2026-10-25 |
| Parasail parasail | parasail-deepseek-v41-flash-1017 | — | — | $0.30 | $1.20 | — | — | — | serverless | 2026-10-24 |
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.
Hub snapshot: 2026-10-25 · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-23 20:12 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.
