Log inSign up
Image
user avatar
DeepInfra
@DeepInfra
Fast ML inference. Run top AI models using a simple API.
Palo Alto
deepinfra.com
Joined February 2023
66
Following
5,550
Followers
RepliesMedia
  • Pinned
    user avatar
    DeepInfra
    @DeepInfra
    Aug 3
    @deepseek_ai's V4 Flash 0731 is rocking the leaderboards, and we're proud to serve it at the best price: $0.09 / $0.18 / $0.018 per 1M 99%+ uptime 46% traffic share on @OpenRouter More optimizations coming 🔥
    user avatar
    DeepInfra
    @DeepInfra
    Jul 31
    🆕DeepSeek-V4-Flash-0731 is now on DeepInfra. The official release of V4-Flash, much stronger agentic capabilities, and it outperforms V4-Pro (Preview) on benchmarks despite a far smaller active parameter count. $0.09 in / $0.18 out. 1M context. 👇
  • user avatar
    DeepInfra
    @DeepInfra
    Aug 3
    @Alibaba_Qwen's Qwen3.8-Max is now live on DeepInfra 🚀 Their 2.4T-parameter MoE flagship — a serious leap in coding and long-horizon agentic work. 256K context · JSON mode · function calling $1.65/M in · $4.95/M out · $0.206/M cached
  • user avatar
    DeepInfra
    @DeepInfra
    Jul 31
    🆕DeepSeek-V4-Flash-0731 is now on DeepInfra. The official release of V4-Flash, much stronger agentic capabilities, and it outperforms V4-Pro (Preview) on benchmarks despite a far smaller active parameter count. $0.09 in / $0.18 out. 1M context. 👇
  • user avatar
    DeepInfra
    @DeepInfra
    Jul 31
    🆕Nano Banana 2 is now on DeepInfra — in two versions 🍌 → nano-banana-2: top-tier image generation + editing, now with Grounding via Google Image Search → nano-banana-2-lite: same great editing, built for speed One OpenAI-compatible API. Try them 👇
  • user avatar
    DeepInfra
    @DeepInfra
    Jul 27
    🆕Claude Opus 5 is now on DeepInfra. Anthropic's most advanced Opus model, built for long-running agents, with real gains in coding and professional work. 1M context. Vision, function calling, JSON. $5 in / $25 out per 1M tokens.