All models

DeepSeek-V4.1-Flash

byDeepSeekDeepSeek· 10 Sep 2026
General purposeMultimodal

Since the latest Artifacts, the whale has not released one, but two new model iterations with a new size (552B-A16B) which also support image inputs. The model has gained a lot of (well-deserved) attention for its performance at a small price. It also carries the MIT license, continuing the trend established by DeepSeek themselves. This is notable in times where licenses seem to become more bespoke and exclusionary. Below is the Pareto of performance vs price on DeepSWE, a favorite agentic software engineering benchmark today, which DeepSeek dominates.

Specs
Params552B
LicenseMIT
Capability · Artificial Analysis
AA Index
39.5
Months Behind Frontier
4.8 mo
claude-opus-4-7
16 Apr 2026
Adoption · Hugging Face
RAM @ 30d
Relative Adoption Metric: 2.74×. Healthy. Measured at 30 days.
Hugging Face Downloads
2.1M
last 30d
3M
all time
HF Likes
4.9K

Relative Adoption Metric contextualizes downloads against the model's size bucket.

Inference · OpenRouter
Tokens/Day
2.2T
7d avg · listed 7/7
Peak Tokens/Day
7.5T
07 Oct 2026
Peak Rank
#1
16 Sep 2026

OpenRouter publishes daily token totals for its ~50 most-served models. Days missing from the chart mean DeepSeek-V4.1-Flash fell below that cutoff — not zero usage. OpenRouter logs usage separately per dated model version and per variant (like ":free"); the hub combines provider variants while keeping separately cataloged releases distinct.

Related Models