DeepSeek V4 Flash 0731

DeepSeek released V4 Flash 0731, an iterative update of its Mixture-of-Experts model with 284B total parameters and 13B active per token, achieving an Artificial Analysis Intelligence Index of 50, up 10 points from the previous version and 6 points above V4 Pro. The model, which now routes the deepseek-chat and reasoner API aliases, shows improved agentic performance (Elo 1559, up from 1189) and a 12-point drop in hallucination rate to 84%, with pricing unchanged at $0.14/$0.28 per 1M tokens. Unsloth has released Dynamic 2.0 GGUFs for local inference, with UD-Q4_K_XL (155GB) and UD-Q8_K_XL (162GB) builds fitting dual DGX Sparks, while full MIT-licensed weights are expected in the coming weeks.

DeepSeek V4 Flash 0731 MoE workstation 284B total, 13B active per token MoE . Same FP4+FP8 hybrid-attention family as V4 Pro: Compressed Sparse Attention CSA + Heavily Compressed Attention HCA across 61 layers, manifold-constrained Hyper-Connections mHC , Muon optimizer, 32T+ pretraining tokens. - Context: 1M native, 384K max output; three modes non-think / think-high / think-max . The 2026-07-31 iterative update of V4 Flash supersedes the April model . Per Artificial Analysis, a 10-point Intelligence Index jump to 50 - 6 points above V4 Pro, 1 behind GLM 5.2 / GPT-5.6 Luna, 7 behind Kimi K3. Agentic Elo 1559 up from 1189 , Terminal-Bench 2.1 79% +17 , Humanity’s Last Exam 37% +5 , GPQA-Diamond 91% +1 , SciCode 50% +5 . Token usage -12%; hallucination rate 84% a 12-point drop ; AA-Omniscience Index -16 +7 . Pricing unchanged at $0.14/$0.28 per 1M in/out cache-hit $0.0028/M, a 98% discount . The model the deepseek-chat / reasoner API aliases now route to retired 2026-07-24 . - Local run 2026-07-31 : Unsloth’s Dynamic 2.0 GGUFs landed local inference. UD-Q4 K XL is a 155GB lossless 4-bit build ~168GB RAM ; UD-Q8 K XL is a 162GB 8-bit full-precision build ~175GB RAM, only 7GB bigger than Q4 because the 13B active experts dominate . Both fit two stacked DGX Sparks 256GB unified via ConnectX-7 or a 192GB+ unified rig; a smaller 3-bit ~110GB RAM that would fit a single 128GB Spark is announced but not yet published. Run via Unsloth or llama.cpp -hf - Ollama only ships the :cloud endpoint, so there is no local Ollama tag. Open weights under MIT full weights expected in the coming weeks per DeepSeek; the Unsloth GGUFs are available now . - 284.0B - 1000k - mit - Jul 2026 Scores Score per dollar 1000 pts per $/M input general score 90 divided by cheapest input price $0.09/M . Higher is better value. See live pricing /models/deepseek-v4-flash-0731/pricing . Run it locally Per-quant memory needs and a static "can you run it?" reference - no rig entry required Can you run it? - reference rigs | Rig | UD-Q4 K XL | UD-Q8 K XL | |---|---|---| | NVIDIA Jetson Orin NX 16GB | | no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing no - cloud cloud-pricing Fit tiers use the same will-it-run logic as the rig finder. For comfortable fits, the badge reflects decode speed: fast =20 t/s, ok 8-20 t/s, slow <8 t/s. t/s is a bandwidth estimate, not a measured benchmark. How can a 24GB GPU run a 744B model? It does not load the model into VRAM. The quantized weights e.g. ~410GB at Q4 sit in system RAM; the GPU only holds the small shared attention and router tensors and accelerates prompt processing. Because GLM 5.2 is a Mixture-of-Experts model, each token activates only ~40B of its 744B params, so llama.cpp streams just those active experts from system RAM to the GPU each token the -cmoe offload path . That makes decode speed bound by system-RAM bandwidth, not GPU bandwidth - single digits on DDR4, which is why these rigs show 3-8 t/s even though they “fit.” A bigger GPU e.g. 2x 3090 keeps more experts resident on-card and raises tok/s; a smaller GPU still runs it but pays the bandwidth tax. A 744B dense model could not run this way - only MoE’s small-active-params trick makes it possible. Aggressive quants 1-2 bit trade accuracy for size - roughly 17% accuracy loss at 2-bit vs full precision, and real long-context work often needs Q5 or Q6 even when lower quants “fit.” Formula estimates here are conservative; real tuned setups can exceed them one HN user reports ~6 tok/s on a 512GB DDR4 + 2x 3090 rig . Download options Or run it in the cloud Live per-provider pricing, throughput and uptime - refreshed 28 days ago via OpenRouter. Click a column to sort. some pricing may be stale - last verified 2026-07-07 | Provider | Type | Input $/M | Output $/M | Cache $/M | Tok/s | Latency | Uptime | Value | |---|---|---|---|---|---|---|---|---| | | API | 0.09 | 0.18 | - | - | - | - | cheapest | Default order: throughput among 95%+ uptime providers, then latency; subscriptions last. Sort by any column. Subscription rows show $/mo in the Value column - per-token columns are "-". Affiliate links are marked sponsored / nofollow. Confirm current pricing on the provider's site before committing. Detailed API pricing page + JSON endpoint → /models/deepseek-v4-flash-0731/pricing See who runs DeepSeek in production → /adoption/deepseek Inference cost over time Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API.