LiveBench Overall ↗2.0 · 2026-10-14 (dataset 2026-09-02)
63.63%Effort: max · single run · equal mean of 7 suites
deepseek-aiListed Aug 21, 2026
DeepSeek-V4-Flash-Vision-Exp
DeepSeek-V4-Flash-Vision-Exp is DeepSeek's experimental multimodal model in the V4-Flash family, adding visual understanding to the V4-Flash architecture. It serves a 1M-token (1,048,576) context window and…
Source description
DeepSeek-V4-Flash-Vision-Exp is DeepSeek's experimental multimodal model in the V4-Flash family, adding visual understanding to the V4-Flash architecture. It serves a 1M-token (1,048,576) context window and supports image input with visual grounding, tool calling, structured/JSON output, and configurable reasoning effort (low/high/max, or disabled).
Context window1M
Max output262.1K
Provider endpoints6
Benchmark results8
Measured performance
All leaderboards ↗Benchmark results
LiveBench Design ↗2.0 · 2026-10-14 (dataset 2026-09-02)
70.7%Effort: max · single run
LiveBench Research ↗2.0 · 2026-10-14 (dataset 2026-09-02)
69.2%Effort: max · single run
LiveBench Browser Use ↗2.0 · 2026-10-14 (dataset 2026-09-02)
64%Effort: max · single run
LiveBench Coding ↗2.0 · 2026-10-14 (dataset 2026-09-02)
37.6%Effort: max · single run
LiveBench Agentic ↗2.0 · 2026-10-14 (dataset 2026-09-02)
39%Effort: max · single run
LiveBench Data Analysis ↗2.0 · 2026-10-14 (dataset 2026-09-02)
79.5%Effort: max · single run
LiveBench Reasoning ↗2.0 · 2026-10-14 (dataset 2026-09-02)
85.4%Effort: max · single run
Access & cost
6 quotesPrices across providers
Your token mix
USD · listed token ratesOpenRouter
Catalog quote- Input
- $0.2156
- Output
- $0.6468
- Cached input
- $0.00686
deepinfra
via Hugging Face- Input
- $0.44
- Output
- $1.32
- Input
- $0.44
- Output
- $1.32
- Input
- $0.2156
- Output
- $0.6468
- Cached input
- $0.00686
Source discount 51%; rates as published
- Input
- $0.44
- Output
- $1.32
- Cached input
- $0.014
- Input
- $0.44
- Output
- $1.32
- Cached input
- $0.028
Model specifications
- Developer
- deepseek-ai
- Modality
- Language
- Context
- 1,048,576 tokens
- Maximum output
- 262,144 tokens
- Weights
- Open weights
- Provider listed
- Aug 21, 2026
- First discovered
- Oct 16, 2026
Exact identities
- openrouter
deepseek/deepseek-v4-flash-vision-exp - huggingface
deepseek-ai/DeepSeek-V4-Flash-Vision-Exp - deepinfra
deepseek-ai/DeepSeek-V4-Flash-Vision-Exp - livebench
deepseek-v4-flash-vision
Sources
Updated Oct 20, 2026
More from deepseek-ai
View all ↗deepseek-aiReleased 3 days ago
DeepSeek-V4.1-Flash-1017
DeepSeek-V4.1-Flash-1017 is a minor update to DeepSeek-V4.1-Flash, delivering substantial improvements across reasoning, coding, agentic tasks, and visual benchmarks. Alongside these gains, the model…
1M contextImage inputTool calling
deepseek-aiListed Aug 12
DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813 is the official release of DeepSeek-V4-Pro, superseding the preview version, with greatly enhanced agentic capabilities and performance improvements that are especially pronounced in…
1M contextTool callingStructured outputNo reported score
No reported score
$0.225in$8.00out/ 1M tokens · USDRelace · via OpenRouter · 30 endpointsrelace/fp4 · quantization=fp4
deepseek-aiListed Jul 31
DeepSeek-V4-Flash-0731
DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version, with substantially enhanced agentic capabilities. DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro…
1M contextTool callingStructured outputNo reported score
No reported score
$0.0086in$0.45out/ 1M tokens · USDRelace · via OpenRouter · 35 endpointsrelace/fp4 · quantization=fp4
deepseek-aiListed Apr 22
DeepSeek-V4-Pro
DeepSeek V4 Pro is an MoE model with 1.6T total parameters (49B active) and a 1M-token context window. It's built for advanced reasoning, coding, and long-running agent tasks, and performs well on…
1M contextTool callingStructured outputNo reported score
No reported score