DeepSeek-V4.1-Flash release with native multimodal support
DeepSeek released V4.1-Flash on its API, with native multimodal support and the model name deepseek-flash.
On Zabio today
deepseek-flashLive. Call it by name, or let auto pick it when it fits the job.Price and detailsdeepseek-v4-flashLive. Call it by name with your Zabio key.Price and detailsdeepseek-v4-flash-vision-expLive. Call it by name, or let auto pick it when it fits the job.Price and detailsdeepseek-v4-proLive. Call it by name, or let auto pick it when it fits the job.Price and details
What changed
DeepSeek released DeepSeek-V4.1-Flash, the smallest model in its new architecture family, with native multimodal visual understanding. It is now available on the DeepSeek API, and developers call it by setting the model name to deepseek-flash. The previous V4 Flash and V4 Flash Vision Exp models have been retired. API prices were reduced with this release. The entry also says DeepSeek V4 Pro continues on the API after September 14, 2026, with the billing method unchanged.
Key details
- GPQA Diamond score: 90.9.
- Codeforces rating: 3471.
- Terminal-Bench 2.1 score: 90.6.
- Architecture: 552B-parameter MoE, with 8B active parameters for input and 16B for output.
- KV cache needs 1/4 the HBM and 1/8 the SSD storage compared with the previous generation.
- Off-peak rates are 50% of peak rates. New pricing takes effect at 04:00 UTC on Sept 10, 2026.
What it means for Zabio customers
If you use V4 Flash or V4 Flash Vision Exp, move to deepseek-flash, because the old names only route temporarily. The entry gives no per-token prices for this model, and the rates are listed on the Models & Pricing page.
Breaking changes and migration
Change the model name to deepseek-flash to call V4.1 Flash. The names deepseek-v4-flash and deepseek-v4-flash-vision-exp are temporarily routed to V4.1 Flash for compatibility.
Source
DeepSeek V4.1 Flash is now available on the DeepSeek API with native multimodal support. Change the model name to deepseek-flash to call the latest V4.1 Flash model.
Original announcement: DeepSeek API change log
How it compares
The models in this story next to models of the same class at a similar price, all available with one Zabio key.
| Model | Input, $/M | Output, $/M | Cached, $/M | Context window | 1,000 calls | Math, hard |
|---|---|---|---|---|---|---|
deepseek-flashDeepSeekIn this story | 0.30 | 1.20 | 0.006 | 1,048,576 | $1.20 | 93.3 |
deepseek-v4-flashDeepSeekIn this story | 0.20 | 0.40 | 0.04 | 1,000,000 | $0.60 | 79.7 |
deepseek-v4-flash-vision-expDeepSeekIn this story | 0.2156 | 0.6468 | 0.00686 | 131,072 | $0.75 | 87.8 |
deepseek-v4-proDeepSeekIn this story | 1.32 | 3.96 | 0.044 | 1,000,000 | $4.62 | 90.7 |
codestral-2508Mistral AI | 0.30 | 0.90 | 0.03 | 128,000 | $1.05 | no data |
gpt-3.5-turboOpenAI | 0.50 | 1.50 | no data | 16,385 | $1.75 | 11.6 |
glm-4.6Z.ai | 0.50 | 2.00 | 0.10 | 131,072 | $2.00 | no data |
Prices are the vendor list prices in USD per million tokens. "1,000 calls" is 1,000 requests of 2,000 input tokens and 500 output tokens each, without caching. Zabio bills the list price plus its service fee: see pricing.
Scores: percent correct for that kind of job, aggregated from public results by Epoch AI and LiveBench, read on 16 November 2026. Epoch AI: CC BY 4.0, LiveBench: Leaderboard results used with attribution; the 2.0 release states no terms of its own (read 2026-11-06); the first version's Mathematics, Language, IF results: Code Apache 2.0 and MIT; leaderboard results used with attribution
Source: DeepSeek API change log (2026-09-10)