The DwarfStar (ds4) Q2 build of DeepSeek-V4.1-Flash-1017 with DeepSeek's refusal direction projected out of the weights. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. - This is antirez/deepseek-v4.1-flash-gguf's `DeepSeek-V4.1-Flash-1017-IQ2_XXS-Q2_K-imatrix.gguf` (revision `9dcee51692e6`) with **only the 122 tensors that write into the residual stream replaced**: attention output (`attn_output_b`), shared-expert down projection, the down projection of all 384 routed experts of every block, and Engram's value projection. Every other byte, the header included, is the original's, so it runs wherever that file runs, in the same memory. - The replaced tensors carry the edit of AtomicChat/DeepSeek-V4.1-Flash-1017-Uncensored-GGUF (one direction, strength 2.0), baked into DeepSeek's weights at full precision and quantized to the original file's types (Q8_0, Q2_K with an importance matrix, F16). After the bake, 99.98% of the edit is in every writer.

Text generation

Input Text→ Output Text

Price not verifiedCheck multimodal and channel pricing separately

View details ↗

CompareSource / docs ↗

DeepSeek-V4.1-Flash-1017-Uncensored-GGUF

AtomicChat

DeepSeek's refusal direction projected out of the released weights, baked into GGUF builds and measured against the original. **`2.67bpw-IQ1_M` costs knowledge.** It answers without refusals like the other build, but at 1 bit the edit takes MMLU from 85.6% to 81.8% (500 questions, p = 0.003; our gate allows 1.5 points), and it did not pass that gate. With thinking on it also goes round more often (3/30 looping traces at low effort and 5/30 at high on JailbreakBench). It is here for machines with about 140 GB for the weights; if `3.58bpw-Q2_K` fits, take that one. Neutral held-out text, 24 chunks of 4096 tokens, against Unsloth's UD-Q8_K_XL (the converter's BF16 + native MXFP4 output of the released checkpoint) as the reference.

Text generation

Input Text→ Output Text

Price not verifiedCheck multimodal and channel pricing separately

View details ↗

CompareSource / docs ↗

DeepSeek-V4-Flash-0731-GGUF

AtomicChat

基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。

Text generation

Input Text→ Output Text

Price not verifiedCheck multimodal and channel pricing separately

View details ↗

CompareSource / docs ↗

embeddinggemma-2-GGUF

AtomicChat

基础档案已收录，详细介绍后续补充。可先查看厂商、分类与来源。

Embeddings & similarity

Input Text→ Output Embedding

Price not verifiedCheck multimodal and channel pricing separately

View details ↗

CompareSource / docs ↗

Flint-GGUF

AtomicChat

Built from RecursiveSI's original weights with our own importance matrix. The calibration corpora behind our builds are public. Built by the team behind Atomic Chat , the app that runs open-weight models locally and offline. Flint is RecursiveSI's 39.8B mixture-of-experts coding model. About 2.5B parameters are active per token. See our measurements below, including how these files compare to other builds of Flint. Generate about 1.5 times faster with the DFlash2 draft model: the command and our speed measurements are in Faster with DFlash2 . Every number below is measured, not estimated. How we measured it is at the bottom, and the raw logs are in the metrics repo so you can check any of it yourself.

Text generation

Input Text→ Output Text

Price not verifiedCheck multimodal and channel pricing separately

View details ↗

CompareSource / docs ↗

COST EXPLORER

Great models. Clear costs.

Cost analysis

Adjust monthly token usage to compare estimated API costs.

Monthly input tokens 10 MMonthly output tokens 2 M1 M = 1 million tokens · USD
Compares base text-token rates within the same channel

Uses collected, comparable USD-per-million-token rates. Snapshots exclude cache discounts, tiers, extra reasoning output and taxes. Check current source prices.

MODEL HANDBOOK

Understand before you choose

Selection guide & FAQ

How should API prices be compared?+

Separate input, output and cache rates, then estimate the same workload. Image, audio and video rates may use items, duration or resolution rather than tokens.

Do open weights mean free commercial use?+

Not necessarily. Check each version’s license, commercial restrictions and policies. Self-hosting also incurs compute and maintenance costs.

Is the top-ranked model the best for me?+

Rankings depend on tasks, versions and methods. Compare accuracy, latency, reliability and costs on your own workload.

Find your ideal AI model.

Explore AI models 84991

Ornith-1.5-9B-Heretic-MultiFormat

Qwable-Claude-Fable-5-Distilled-NPU2

Qwen3.8-27B-Heretic-MTP-GGUF

Qwen3.8-27B-NPU2

Qwen3.8-35B-A3B-Distill-GGUF

Qwen3.8-9B-Distill-GGUF

Qwen3.8-9B-Distill-uncensored-heretic-Q4_K_M-GGUF

Qwen3.8-Distilled-2B-NPU2

Qwen3.8-Distilled-4B-NPU2

Qwen3.8-Distilled-9B-NPU2

Qwen3.8-Distilled-9B-NPU2-GGUF

Qwen3.8-Flash-Next-APEX-GGUF

Qwen3.8-Flash-Next-Prune11264-124B-GGUF

Qwen3.8-Whittle-MoE-27B-A17.8B

Qwen3.8-Whittle-MoE-27B-A17.8B-GGUF

Qwopus3.5-9B-Coder-NPU2

Beam-1.0-501B-A23B-GGUF

d1-3B-GGUF

d1-omni-600M-GGUF

DeepSeek-V4.1-Flash-1017-Uncensored-DwarfStar-GGUF

DeepSeek-V4.1-Flash-1017-Uncensored-GGUF

DeepSeek-V4-Flash-0731-GGUF

embeddinggemma-2-GGUF

Flint-GGUF

No matching models

Great models. Clear costs.

Understand before you choose

Follow trends. Find tools. Explore models.