DeepSeek-V4-Flash Local Inference Hits 120 tokens/s via Unsloth
danielhanchen · x · 2026-08-06
UnslothAI announced that with the DSpark tool, local running speed of DeepSeek-V4-Flash-0731 GGUFs can be increased by 1.4 to 2× without accuracy change. The optimized model can reach up to 120 tokens/s generation speed.
Related event: DeepSeek V4 Flash Benchmarks Leak, Outperforming Pro(3 posts)→
More from Infra
- Same 1,000 Cloudflare Durable Objects: $4,050/month cut to $22 with one trick — threepointone · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- Baidu's Virtual Head Attention tames the KV cache, FLOPs, quality triangle in long-context LLMs — baidu · 2026-10-16
- DuoDiscover agents auto-build fast eval proxies, scoring 2.68x TPU kernel speedup — berkeley_ai · 2026-10-16
- Google Raises 2026 Capital Spending to Up to $205 Billion — emmanuelvivier · 2026-10-16
- Google open-sources AX, a Kubernetes-style orchestrator for billions of AI agent tasks — thisdudelikesAI · 2026-10-16