{"data":{"capabilities":{"chat":null,"reasoning":null,"tools":null,"vision":null},"description":"Observed LocalMaxxing leaderboard run. Evidence for compatibility, not an executable launch contract.","engine":{"graph_mode":null,"name":"llama.cpp","version":"69 (fork ds4-longctx @ e6718011c)"},"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"},"serving.max_concurrency":{"provenance":{"captured_at":"2026-08-31T23:03:15Z","sources":[{"captured_at":"2026-08-31T23:03:15Z","kind":"registry-derived","url":"https://www.localmaxxing.com/en/runs/cmsi0l2sc01lepp019uelu1ga"}]},"reason":"server-capacity-derived-from-source-evidence","state":"known"}},"hardware_count":4,"hardware_id":"rtx-3090-24gb","id":"deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4","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/cmsi0l2sc01lepp019uelu1ga"}],"state":"none"},"kind":"reference","run_id":"cmsi0l2sc01lepp019uelu1ga","source":"localmaxxing","url":"https://www.localmaxxing.com/en/runs/cmsi0l2sc01lepp019uelu1ga"},"metadata":{"localmaxxing":{"backend":"cuda","batch_size":1,"hardware_label":"RTX 3090","notes":"DeepSeek-V4-Flash-0731 (304B MoE, Unsloth UD-IQ2_M, ~90GB GGUF) on 4x RTX 3090 24GB (96GB total, ~90.3GiB used at 130K ctx). Engine: llama.cpp fork with DeepSeek-V4-Flash optimizations - github.com/alesha-pro/llama.cpp branch ds4-longctx @ e6718011c. Launcher: scripts/ds4-prod-serve.sh (full env flag set in commandSnippet). Custom work vs upstream: prefill CUDA graphs, sparse FA + indexer skip, fused MoE (GLU fuse + tiling + resident experts), radix top-k, register-based MLA head-compression sinkhorn kernel, 2-pass fp8 KV quantize kernel, sticky graph allocator, startup warm pass (moves allocator climb out of first request). GPU tuning: PL 270W/card + locked SM clocks (scripts/ds4-gpu-clocks.sh). Full story: DS4HANDOFF.md in the repo. Measured 2026-08-06, server fully warm. This run: fresh 32768-token prompt -> prefill 1745.4 t/s (TTFT 18.8s, ~265W/card), then decode 41.07 t/s at 32K depth (~210W/card). Same config decodes 44.6 t/s at short context.","observed_command":"CUDA_VISIBLE_DEVICES=0,1,2,3 GGML_CUDA_P2P=1 DSV4_CONSTANT_SHAPE=1 DSV4_SPARSE_FA=1 DSV4_FA_UNION=1 DSV4_IDX_SKIP=1 DSV4_MOE_TILE=1 DSV4_MOE_RESIDENT=1 DSV4_GLU_FUSE=1 DSV4_MOE_FUSE=1 DSV4_DECODE_FUSED_IDX=1 DSV4_DECODE_RADIX_TOPK=1 DSV4_PREFILL_RADIX_TOPK=1 DSV4_MMVQ_SMALLK=1 DSV4_STABLE_TOPO=1 GGML_GALLOC_STICKY=1 DSV4_PREFILL_GRAPHS=1 DSV4_AGENT_CKPT_TAIL=1 llama-server -m DeepSeek-V4-Flash-0731-UD-IQ2_M-00001-of-00003.gguf -ngl 999 -ts 1,1,0.95,1.05 --split-mode layer --flash-attn on --no-repack --ctx-size 131072 --batch-size 8192 --ubatch-size 384 --ctx-checkpoints 1 -t 8 --parallel 1 --jinja --reasoning-budget 1024 --reasoning-budget-message 'Okay, that is enough thinking. Let me get to work.'","run_id":"cmsi0l2sc01lepp019uelu1ga","tokenized":{"arguments":["llama-server","-m","DeepSeek-V4-Flash-0731-UD-IQ2_M-00001-of-00003.gguf","-ngl","999","-ts","1,1,0.95,1.05","--split-mode","layer","--flash-attn","on","--no-repack","--ctx-size","131072","--batch-size","8192","--ubatch-size","384","--ctx-checkpoints","1","-t","8","--parallel","1","--jinja","--reasoning-budget","1024","--reasoning-budget-message","Okay, that is enough thinking. Let me get to work."],"environment":{"CUDA_VISIBLE_DEVICES":"0,1,2,3","DSV4_AGENT_CKPT_TAIL":"1","DSV4_CONSTANT_SHAPE":"1","DSV4_DECODE_FUSED_IDX":"1","DSV4_DECODE_RADIX_TOPK":"1","DSV4_FA_UNION":"1","DSV4_GLU_FUSE":"1","DSV4_IDX_SKIP":"1","DSV4_MMVQ_SMALLK":"1","DSV4_MOE_FUSE":"1","DSV4_MOE_RESIDENT":"1","DSV4_MOE_TILE":"1","DSV4_PREFILL_GRAPHS":"1","DSV4_PREFILL_RADIX_TOPK":"1","DSV4_SPARSE_FA":"1","DSV4_STABLE_TOPO":"1","GGML_CUDA_P2P":"1","GGML_GALLOC_STICKY":"1"},"fidelity":"faithful"}}},"model_instance_id":"unsloth-deepseek-v4-flash-0731-gguf--ud-iq2-m","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/cmsi0l2sc01lepp019uelu1ga"}]},"recipe_source":"localmaxxing","schema_version":"local-ai-registry/v1","serving":{"kv_cache_tokens":null,"max_concurrency":1,"max_context_tokens":131072,"tensor_parallel":4},"speed_sweep_ids":["deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4-sweep"],"status":"candidate","huggingface":{"link_type":"repository","provenance":{"captured_at":"2026-08-30T09:10:02Z","sources":[{"captured_at":"2026-08-30T09:10:02Z","kind":"huggingface-api","url":"https://huggingface.co/api/models/unsloth/DeepSeek-V4-Flash-0731-GGUF"}]},"reason":"hf-api-confirmed-public","repository":"unsloth/DeepSeek-V4-Flash-0731-GGUF","status":"known","url":"https://huggingface.co/unsloth/DeepSeek-V4-Flash-0731-GGUF"},"registry":{"launchable":false,"speed_evidence":{"available":true,"count":1,"speed_sweep_ids":["deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4-sweep"],"detail_urls":["/api/v1/speed-sweep/deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4-sweep"]},"runtime":"reference"},"relationships":{"hardware":{"api":"/api/v1/hardware/rtx-3090-24gb","href":"/hardware/rtx-3090-24gb","id":"rtx-3090-24gb","name":"GeForce RTX 3090"},"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/unsloth-deepseek-v4-flash-0731-gguf--ud-iq2-m","href":"/model-instances/unsloth-deepseek-v4-flash-0731-gguf--ud-iq2-m","id":"unsloth-deepseek-v4-flash-0731-gguf--ud-iq2-m","name":"unsloth/DeepSeek-V4-Flash-0731-GGUF"},"speed_sweep":[{"api":"/api/v1/speed-sweep/deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4-sweep","href":"/speed-sweep/deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4-sweep","id":"deepseek-v4-flash-0731-ud-iq2-m-rtx-3090-24gb-llama-cpp-tp4-sweep"}]}},"meta":{"source":"registry"}}