{"name":"Nemotron-3 Nano Omni 30B-A3B","arena":{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4","name":"Nemotron-3 Nano Omni 30B-A3B","vram":21,"params":30,"vendor":"nvidia","quality":{"coding":null,"science":null,"knowledge":null},"contextK":256,"baseModelId":"nemotron-3-nano-omni-30b-a3b-reasoning"},"notes":"Model-ID: nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4. Container: vllm/vllm-openai:v0.26.0-aarch64-cu129-ubuntu2404. llama-benchy 0.4.0. Bronrun: 2026-08-07T12:24:23+02:00. KV-cache auto; max_model_len 131072; gpu_memory_utilization 0,9. Prefix caching uit; async scheduling uit. Extra serverflags: --trust-remote-code --max-num-seqs 256 --max-num-batched-tokens 8192 --video-pruning-rate 0.5 --reasoning-parser nemotron_v3 --enable-auto-tool-choice --tool-call-parser qwen3_coder --moe-backend marlin. 10 van de 11 tests hebben een meting die volgens de huidige Arena-regels meetelt. Bij deze historische run is geen sanity-check vastgelegd. Dat bewijs ontbreekt, ook wanneer de snelheidsmeting volgens de huidige regels meetelt. Afgebroken: 11-rate-sweep.","order":250,"engine":"vLLM v0.26.0","verdict":"Gemeten op 2026-08-07 met vLLM v0.26.0. In de chattest met tien gelijktijdige verzoeken haalt dit profiel 23,94 tokens/s per gebruiker, met gemiddeld 1,48 seconden tot het eerste token. Bij 25k context stijgt die gemiddelde wachttijd naar 17,38 seconden. Bij deze historische run is geen sanity-check vastgelegd. Dat bewijs ontbreekt, ook wanneer de snelheidsmeting volgens de huidige regels meetelt. De capaciteitsmeting is afgebroken; daarmee is geen volledig capaciteitsplafond vastgesteld.","hardware":"DGX Spark, NVIDIA GB10, 128 GB unified memory","modelUrl":"https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4","provider":"nvidia","precision":"NVFP4","parameters":"30B-A3B","provenance":{"modelId":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4","runPath":"results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4","validity":{"tests":{"01-chat":"complete","02-rag-8k":"complete","08-sharegpt":"complete","09-reasoning":"complete","04-multi-turn":"complete","11-rate-sweep":"aborted","03-long-output":"complete","05-big-context":"complete","10-monday-peak":"complete","07-office-baseline":"complete","06-concurrency-stress":"complete"},"sanity":"not_recorded"},"sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","servedName":"nemotron-3-nano-omni-30b-nvfp4","generatedAt":"2026-08-07T12:24:23+02:00","runRevision":"a948cec23b054c62fef595bad3cdb284066922ac","vllmVersion":"0.26.0","suiteVersion":"2026-08","repositoryUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks","contractVersion":1,"contractRevision":"a948cec23b054c62fef595bad3cdb284066922ac","commandProvenance":"reconstructed_from_raw_artifacts","llamaBenchyVersion":"0.4.0"},"architecture":"MoE","measuredContextK":128,"benchmarkResults":[{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-chat-1k","date":"2026-08-07","runs":3,"metrics":{"ttfrSec":{"mean":1.475,"stddev":0.299},"decodeTokensPerSecondTotal":{"mean":218.91,"stddev":1.61},"prefillTokensPerSecondTotal":{"mean":5744.17,"stddev":43.07},"decodeTokensPerSecondPerUser":{"mean":23.94,"stddev":1.77}},"sourceRun":"Test 01","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","workloadId":"chat-1k","sourceLabel":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"01-chat","prefixCaching":false},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-rag-8k","date":"2026-08-07","runs":3,"metrics":{"ttfrSec":{"mean":5.974,"stddev":2.92},"decodeTokensPerSecondTotal":{"mean":152.9,"stddev":4.75},"prefillTokensPerSecondTotal":{"mean":7383.21,"stddev":110.01},"decodeTokensPerSecondPerUser":{"mean":19.73,"stddev":3.2}},"sourceRun":"Test 02","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","workloadId":"rag-8k","sourceLabel":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"02-rag-8k","prefixCaching":false},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-long-output-agents","date":"2026-08-07","runs":3,"metrics":{"ttfrSec":{"mean":0.821,"stddev":0.075},"decodeTokensPerSecondTotal":{"mean":159.47,"stddev":5.22},"prefillTokensPerSecondTotal":{"mean":2774.59,"stddev":67.52},"decodeTokensPerSecondPerUser":{"mean":29.15,"stddev":4.91}},"sourceRun":"Test 03","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","workloadId":"long-output-agents","sourceLabel":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"03-long-output","prefixCaching":false},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-multi-turn-office","date":"2026-08-07","runs":3,"metrics":{"ttfrSec":{"mean":2.143,"stddev":0.714},"decodeTokensPerSecondTotal":{"mean":206.65,"stddev":4.36},"prefillTokensPerSecondTotal":{"mean":6571.29,"stddev":243.06},"decodeTokensPerSecondPerUser":{"mean":23.6,"stddev":1.43}},"sourceRun":"Test 04","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","workloadId":"multi-turn-office","sourceLabel":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"04-multi-turn","prefixCaching":false},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-large-context-25k","date":"2026-08-07","runs":3,"metrics":{"ttfrSec":{"mean":17.378,"stddev":8.737},"decodeTokensPerSecondTotal":{"mean":62.38,"stddev":0.24},"prefillTokensPerSecondTotal":{"mean":7304.27,"stddev":9.33},"decodeTokensPerSecondPerUser":{"mean":11.66,"stddev":4.93}},"sourceRun":"Test 05","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","workloadId":"large-context-25k","sourceLabel":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"05-big-context","prefixCaching":false},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-concurrency-stress-25k","date":"2026-08-07","runs":3,"metrics":{"ttfrSec":{"mean":32.651,"stddev":17.602},"decodeTokensPerSecondTotal":{"mean":67.62,"stddev":1.52},"prefillTokensPerSecondTotal":{"mean":7326.03,"stddev":9.61},"decodeTokensPerSecondPerUser":{"mean":7.21,"stddev":3.8}},"sourceRun":"Test 06","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","workloadId":"concurrency-stress-25k","sourceLabel":"nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"06-concurrency-stress","prefixCaching":false}],"openLoopResults":[{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-office-random-4k","date":"2026-08-07","metrics":{"e2eMs":{"p50":20669.67,"p90":39104.8,"p95":44226.04,"p99":51212.13,"mean":21465.98},"tpotMs":{"p50":41.66,"p90":58.11,"p95":65.43,"p99":72.15,"mean":42.46},"ttftMs":{"p50":1088.79,"p90":1880.99,"p95":2324.28,"p99":2551.33,"mean":1193.55}},"sourceRun":"Test 07","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","scenarioId":"office-random-4k","achievedRps":0.291,"sourceStatus":"complete","sourceTestId":"07-office-baseline","configuredRps":0.3,"prefixCaching":false,"totalRequests":200,"successfulRequests":200,"totalTokenThroughput":1328.94,"peakConcurrentRequests":15},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-sharegpt-replay","date":"2026-08-07","metrics":{"e2eMs":{"p50":2807.2,"p90":11832.63,"p95":13749.3,"p99":19256.99,"mean":4716.34},"tpotMs":{"p50":20.87,"p90":32.96,"p95":34.88,"p99":40.93,"mean":22.52},"ttftMs":{"p50":130,"p90":569.98,"p95":615.53,"p99":967.82,"mean":205.58}},"sourceRun":"Test 08","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","scenarioId":"sharegpt-replay","achievedRps":0.3,"sourceStatus":"complete","sourceTestId":"08-sharegpt","configuredRps":0.3,"prefixCaching":false,"totalRequests":250,"successfulRequests":250,"totalTokenThroughput":131.28,"peakConcurrentRequests":9},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-reasoning","date":"2026-08-07","metrics":{"e2eMs":{"p50":183733.74,"p90":268046.91,"p95":302423.31,"p99":318083.89,"mean":176543.11},"tpotMs":{"p50":43.42,"p90":48.88,"p95":49.37,"p99":50.76,"mean":42.66},"ttftMs":{"p50":587.71,"p90":727.31,"p95":824.42,"p99":967.31,"mean":531.45}},"sourceRun":"Test 09","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","scenarioId":"reasoning","achievedRps":0.106,"sourceStatus":"complete","sourceTestId":"09-reasoning","configuredRps":0.2,"prefixCaching":false,"totalRequests":50,"successfulRequests":50,"totalTokenThroughput":564.38,"peakConcurrentRequests":35},{"id":"nemotron-3-nano-omni-30b-a3b-reasoning-nvfp4-monday-peak-random-4k","date":"2026-08-07","metrics":{"e2eMs":{"p50":42957.58,"p90":83178.98,"p95":86065.02,"p99":90925.42,"mean":46346.14},"tpotMs":{"p50":92.92,"p90":100.54,"p95":107.38,"p99":123.42,"mean":91.8},"ttftMs":{"p50":1005.05,"p90":1725.2,"p95":2134.34,"p99":4441.22,"mean":1168.76}},"sourceRun":"Test 10","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","scenarioId":"monday-peak-random-4k","achievedRps":0.521,"sourceStatus":"complete","sourceTestId":"10-monday-peak","configuredRps":1.5,"prefixCaching":false,"totalRequests":300,"successfulRequests":300,"totalTokenThroughput":2350.14,"peakConcurrentRequests":27}],"rateSweep":{"date":"2026-08-07","capacity":{"slo2000":{"outputTps":48.4,"ttftP95Ms":1704.8,"achievedRps":0.099,"configuredRps":0.1,"peakConcurrent":5},"slo5000":{"outputTps":139.1,"ttftP95Ms":3177.3,"achievedRps":0.283,"configuredRps":0.3,"peakConcurrent":18},"slo10000":{"outputTps":139.1,"ttftP95Ms":3177.3,"achievedRps":0.283,"configuredRps":0.3,"peakConcurrent":18}},"sourceRun":"Test 11","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-3/nemotron-3-nano-omni-30b-a3b-reasoning/nvfp4/","abortedAtRps":0.5,"queueKneeRps":null,"sourceStatus":"aborted","sourceTestId":"11-rate-sweep","totalRequests":300,"successfulRequests":300}}