DeepSeek-V4-Flash-Vision-Exp is an experimental multimodal visual understanding model launched by DeepSeek, officially available on the DeepSeek API platform as of August 21, 2026. Built upon the DeepSeek-V4-Flash architecture (284B total parameters, 13B activated MoE during inference), this model matches the pure-text capabilities of the official DeepSeek-V4-Flash while achieving significant leaps in visual understanding Agent performance, approaching the level of Opus-4.8.
Back to Models
-$0.22-0.44/ M tokens$0.66-1.32/ M tokensRead:0.007-0.014/ M tokensWrite:-/ M tokens1M--
99.2%$0.22-0.44/ M tokens$0.66-1.32/ M tokensRead:0.007-0.014/ M tokensWrite:-/ M tokens1M56.6s105tps
Providers
Route requests across multiple providers. Copy a provider slug to set your preference.
ProviderCache Hit RateContext
DeepSeekSunset
deepseek
TencentCloudSunset
tencent-cloud
Uptime
24hoursDirect request success rate on AI Gateway and per-provider.
Sep 17, 07:59
TencentCloud100.00%
Throughput
24hoursP50 throughput on live AI Gateway traffic, in tokens per second (TPS).
Sep 17, 07:00
TencentCloud105 tps
Latency
24hoursP50 time to first token (TTFT) on live AI Gateway traffic, in milliseconds.
Sep 17, 07:00
TencentCloud56.55s
Activity
Token volume and request traffic to this model over time.
Token Consumption
Sep 16
DeepSeek-
TencentCloud464.3M
Token volume broken down by the provider that served each request. Larger segments indicate providers handling more of this model's traffic.
Benchmarks
Scores on standardized evaluations. Higher percentages are better — and rank percentile shows
Metrics sourced fromArtificial Analysis
Artificial Analysis
Apps
Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for. View All
Token Ranking
No app usage is available for this period.
Related Models
More models from DeepSeek