{"data":{"capabilities":{"chat":null,"reasoning":null,"tools":null,"vision":null},"description":"Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.","draft_launch":{"accelerator_backend":"nvidia","arguments":["--model","deepseek-ai/DeepSeek-V4-Flash-0731","--tensor-parallel-size","2","--host","0.0.0.0","--port","8000","--max-model-len","262144"],"container_port":8000,"entrypoint":null,"environment":{},"host_port":8000,"image":"vllm/vllm-openai@sha256:0a51ea5b4ae2dc5d81890e5173f54203d2a3ae0cfffe51b8fd2afd4391bfd967","ipc":"host","kind":"docker","mounts":[{"read_only":false,"source":"~/.cache/huggingface","target":"/root/.cache/huggingface"}],"shm_size":"16g","synthesized":{"generated_at":"2026-08-31T22:12:17Z","image_provenance":"deepseek-fp8-rtx-pro-6000-blackwell-96gb-vllm-tp1","template":"vllm-openai-v1"}},"engine":{"graph_mode":null,"name":"vllm","version":"0.11.2.dev279+eldritch.enlightenment.v45c1582.b12xf3686b5.cu132.20260704"},"facts":{"capabilities.chat":{"provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"registry-enrichment","url":"https://github.com/0xSero/local-ai-registry"}]},"reason":"capability-not-verified","state":"unknown"},"capabilities.reasoning":{"provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"registry-enrichment","url":"https://github.com/0xSero/local-ai-registry"}]},"reason":"capability-not-verified","state":"unknown"},"capabilities.tools":{"provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"registry-enrichment","url":"https://github.com/0xSero/local-ai-registry"}]},"reason":"capability-not-verified","state":"unknown"},"capabilities.vision":{"provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"registry-enrichment","url":"https://github.com/0xSero/local-ai-registry"}]},"reason":"capability-not-verified","state":"unknown"},"engine.graph_mode":{"provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"registry-enrichment","url":"https://github.com/0xSero/local-ai-registry"}]},"reason":"runtime-detail-not-published","state":"unknown"},"serving.kv_cache_tokens":{"provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"registry-enrichment","url":"https://github.com/0xSero/local-ai-registry"}]},"reason":"kv-cache-capacity-not-published","state":"unknown"}},"hardware_count":2,"hardware_id":"rtx-pro-6000-blackwell-96gb","id":"deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2","launch":{"container":{"captured_at":"2026-08-30T09:10:02Z","compose_file":null,"digest":null,"image":null,"reason":"reference-only-launch","runtime":null,"source":[{"captured_at":"2026-08-30T09:10:02Z","kind":"recipe-launch","url":"https://www.localmaxxing.com/en/runs/cmsb828qn0061qd01l1fsm6pa"}],"state":"none"},"kind":"reference","run_id":"cmsb828qn0061qd01l1fsm6pa","source":"localmaxxing","url":"https://www.localmaxxing.com/en/runs/cmsb828qn0061qd01l1fsm6pa"},"metadata":{"localmaxxing":{"backend":"cuda","hardware_label":"RTX PRO 6000 Blackwell","notes":"First DeepSeek-V4-Flash-0731 (re-post-trained, same 284B/13B arch) on 2x RTX PRO 6000 Blackwell (SM120). Backend: FLASHINFER_MLA_SPARSE_DSV4 attn + flashinfer_cutlass MoE (voipmonitor 'lucifer-cutlass'), NOT b12x. Native DSpark draft head, K5 (dspark_block_size=5 is baked into config.json = hard ceiling).\n\nWORKLOAD: tokSOut 276.0 is single-stream CODE generation, to stay comparable with the other 2x RTX PRO 6000 DSpark runs here. Same runs gave prose 149.8 and math 285.8 t/s - prose is ~46% below code on this stack, so workload choice dominates any hardware difference.\n\nMETHOD: greedy (temp 0), thinking=false, two-point decode (64 vs 896 max_tokens on a prefix-warmed identical prompt) so prefill cancels. Mean of 3 runs; spread ~7%. Code prompt = 'commented B-tree with insert/search/delete plus a test harness'. tokSPrefill measured separately as TTFT on a 107,016-token prompt (10.24 s). Short-prompt TTFT 63 ms. peakVramGb is PER CARD, measured: 96799/96917 MiB of 97887 MiB at gpu-memory-utilization 0.93.\n\nVS PREDECESSOR: re-benched the older DeepSeek-V4-Flash-DSpark checkpoint on this exact stack, same session, fresh-vs-fresh, 3 runs each: code 262.6 / prose 155.0 / math 274.9. So 0731 is ~5% faster on code, ~4% on math, ~3% slower on prose - net a wash. 0731 ran on a COLD JIT/autotune cache vs the predecessor's warm one, so it's if anything understated. Spec acceptance unchanged: 48.6% (2.43 per 5-token draft) vs 49.9% (2.50).\n\nCONTEXT: server configured at the full 262144 (not 2048), and the decode figure above is with that KV budget allocated. Cold prefill of a 211,722-token prompt takes 34.6 s (~6.1k tok/s - prefill throughput degrades with length; the 10.4k tok/s above is at 107k). A repeat of the same prompt returns in 0.9 s on the prefix cache. Config allows 1M, but fp8 weights (~81 GiB/card at TP2) cap usable KV at 256k on 2 cards.\n\nHOST: ASRock Rack TURIND8UD-2T, Turin ES 64c/128t, 125 GB, Deb13, drv 590.48.01, TP2 over PCIe Gen5 P2P (no NVLink), 600 W/card.","observed_command":"vllm serve deepseek-ai/DeepSeek-V4-Flash-0731 --tensor-parallel-size 2 --decode-context-parallel-size 1 --kv-cache-dtype fp8 --block-size 256 --gpu-memory-utilization 0.93 --max-model-len 262144 --max-num-seqs 64 --max-num-batched-tokens 8192 --max-cudagraph-capture-size 512 --compilation-config '{\"cudagraph_mode\":\"FULL_AND_PIECEWISE\",\"custom_ops\":[\"all\"]}' --async-scheduling --enable-chunked-prefill --enable-flashinfer-autotune --tokenizer-mode deepseek_v4 --speculative-config '{\"method\":\"dspark\",\"num_speculative_tokens\":5,\"draft_sample_method\":\"probabilistic\"}' --attention-backend FLASHINFER_MLA_SPARSE_DSV4 --kernel-config.moe_backend flashinfer_cutlass --enable-prefix-caching","run_id":"cmsb828qn0061qd01l1fsm6pa","tokenized":{"arguments":["vllm","serve","deepseek-ai/DeepSeek-V4-Flash-0731","--tensor-parallel-size","2","--decode-context-parallel-size","1","--kv-cache-dtype","fp8","--block-size","256","--gpu-memory-utilization","0.93","--max-model-len","262144","--max-num-seqs","64","--max-num-batched-tokens","8192","--max-cudagraph-capture-size","512","--compilation-config","{\"cudagraph_mode\":\"FULL_AND_PIECEWISE\",\"custom_ops\":[\"all\"]}","--async-scheduling","--enable-chunked-prefill","--enable-flashinfer-autotune","--tokenizer-mode","deepseek_v4","--speculative-config","{\"method\":\"dspark\",\"num_speculative_tokens\":5,\"draft_sample_method\":\"probabilistic\"}","--attention-backend","FLASHINFER_MLA_SPARSE_DSV4","--kernel-config.moe_backend","flashinfer_cutlass","--enable-prefix-caching"],"fidelity":"faithful"}}},"model_instance_id":"deepseek-ai-deepseek-v4-flash-0731--fp8-e4m3","provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"normalized-recipe","url":"https://www.localmaxxing.com/en/runs/cmsb828qn0061qd01l1fsm6pa"}]},"recipe_source":"localmaxxing","schema_version":"local-ai-registry/v1","serving":{"kv_cache_tokens":null,"max_concurrency":1,"max_context_tokens":262144,"tensor_parallel":2},"speed_sweep_ids":["deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2-sweep"],"status":"candidate","huggingface":{"link_type":"repository","provenance":{"captured_at":"2026-08-27T06:04:13.773Z","sources":[{"captured_at":"2026-08-27T06:04:13.773Z","kind":"huggingface-api","url":"https://huggingface.co/api/models/deepseek-ai/DeepSeek-V4-Flash-0731"}]},"reason":"hf-api-confirmed-public","repository":"deepseek-ai/DeepSeek-V4-Flash-0731","status":"known","url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731"},"registry":{"launchable":false,"speed_evidence":{"available":true,"count":1,"speed_sweep_ids":["deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2-sweep"],"detail_urls":["/api/v1/speed-sweep/deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2-sweep"]},"runtime":"reference"},"relationships":{"hardware":{"api":"/api/v1/hardware/rtx-pro-6000-blackwell-96gb","href":"/hardware/rtx-pro-6000-blackwell-96gb","id":"rtx-pro-6000-blackwell-96gb","name":"RTX PRO 6000 Blackwell"},"model":{"api":"/api/v1/models/deepseek-v4-flash-0731","href":"/models/deepseek-v4-flash-0731","id":"deepseek-v4-flash-0731","name":"DeepSeek-V4-Flash-0731"},"model_instance":{"api":"/api/v1/model-instances/deepseek-ai-deepseek-v4-flash-0731--fp8-e4m3","href":"/model-instances/deepseek-ai-deepseek-v4-flash-0731--fp8-e4m3","id":"deepseek-ai-deepseek-v4-flash-0731--fp8-e4m3","name":"deepseek-ai/DeepSeek-V4-Flash-0731"},"speed_sweep":[{"api":"/api/v1/speed-sweep/deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2-sweep","href":"/speed-sweep/deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2-sweep","id":"deepseek-v4-flash-0731-fp8-e4m3-rtx-pro-6000-blackwell-96gb-vllm-tp2-sweep"}]}},"meta":{"source":"registry"}}