{"name":"Nemotron-Cascade-2 30B-A3B","arena":{"id":"nemotron-cascade-2-30b-a3b-bf16","name":"Nemotron-Cascade-2 30B-A3B","vram":60,"params":30,"vendor":"nvidia","quality":{"coding":{"bench":"LiveCodeBench v6","value":87.2},"science":{"bench":"GPQA-Diamond","value":76.1},"knowledge":{"bench":"MMLU-Pro","value":79.8}},"contextK":1000},"notes":"Model-ID: nvidia/Nemotron-Cascade-2-30B-A3B. Container: vllm/vllm-openai:v0.26.0-aarch64-cu129-ubuntu2404. llama-benchy 0.4.0. Bronrun: 2026-08-13T08:57:02+02:00. KV-cache fp8; max_model_len 131072; gpu_memory_utilization 0,9. Prefix caching uit; async scheduling aan. Extra serverflags: --trust-remote-code --max-num-batched-tokens 8192. 11 van de 11 tests hebben een meting die volgens de huidige Arena-regels meetelt. De sanity-check is geslaagd. Dit is een controle op een kort gegenereerd antwoord, geen inhoudelijke evaluatie van de benchmarktaken.","order":270,"engine":"vLLM v0.26.0","verdict":"Gemeten op 2026-08-09 tot 2026-08-13 met vLLM v0.26.0. In de chattest met tien gelijktijdige verzoeken haalt dit profiel 7,59 tokens/s per gebruiker, met gemiddeld 1,5 seconden tot het eerste token. Bij 25k context stijgt die gemiddelde wachttijd naar 19,06 seconden. Deze runs meten snelheid en wachttijd, niet de kwaliteit van de antwoorden.","hardware":"DGX Spark, NVIDIA GB10, 128 GB unified memory","modelUrl":"https://huggingface.co/nvidia/Nemotron-Cascade-2-30B-A3B","provider":"nvidia","precision":"BF16","parameters":"30B-A3B","provenance":{"modelId":"nvidia/Nemotron-Cascade-2-30B-A3B","runPath":"results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16","validity":{"tests":{"01-chat":"complete","02-rag-8k":"complete","08-sharegpt":"complete","09-reasoning":"complete","04-multi-turn":"complete","11-rate-sweep":"complete","03-long-output":"complete","05-big-context":"complete","10-monday-peak":"complete","07-office-baseline":"complete","06-concurrency-stress":"complete"},"sanity":"passed"},"sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","servedName":"nemotron-cascade-2-30b-bf16","generatedAt":"2026-08-13T08:57:02+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-cascade-2-30b-a3b-bf16-chat-1k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":1.497,"stddev":0.505},"decodeTokensPerSecondTotal":{"mean":75.25,"stddev":0.39},"prefillTokensPerSecondTotal":{"mean":5113.97,"stddev":12.1},"decodeTokensPerSecondPerUser":{"mean":7.59,"stddev":0.05}},"sourceRun":"Test 01","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","workloadId":"chat-1k","sourceLabel":"nvidia/Nemotron-Cascade-2-30B-A3B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"01-chat","prefixCaching":false},{"id":"nemotron-cascade-2-30b-a3b-bf16-rag-8k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":6.215,"stddev":3.195},"decodeTokensPerSecondTotal":{"mean":67.24,"stddev":0.44},"prefillTokensPerSecondTotal":{"mean":6915.74,"stddev":11.56},"decodeTokensPerSecondPerUser":{"mean":7.25,"stddev":0.31}},"sourceRun":"Test 02","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","workloadId":"rag-8k","sourceLabel":"nvidia/Nemotron-Cascade-2-30B-A3B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"02-rag-8k","prefixCaching":false},{"id":"nemotron-cascade-2-30b-a3b-bf16-long-output-agents","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":0.673,"stddev":0.176},"decodeTokensPerSecondTotal":{"mean":67.88,"stddev":1.43},"prefillTokensPerSecondTotal":{"mean":2926.11,"stddev":50.84},"decodeTokensPerSecondPerUser":{"mean":8.53,"stddev":0.99}},"sourceRun":"Test 03","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","workloadId":"long-output-agents","sourceLabel":"nvidia/Nemotron-Cascade-2-30B-A3B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"03-long-output","prefixCaching":false},{"id":"nemotron-cascade-2-30b-a3b-bf16-multi-turn-office","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":2.127,"stddev":0.777},"decodeTokensPerSecondTotal":{"mean":73.92,"stddev":1.25},"prefillTokensPerSecondTotal":{"mean":6099.16,"stddev":12.68},"decodeTokensPerSecondPerUser":{"mean":7.58,"stddev":0.15}},"sourceRun":"Test 04","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","workloadId":"multi-turn-office","sourceLabel":"nvidia/Nemotron-Cascade-2-30B-A3B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"04-multi-turn","prefixCaching":false},{"id":"nemotron-cascade-2-30b-a3b-bf16-large-context-25k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":19.063,"stddev":9.652},"decodeTokensPerSecondTotal":{"mean":41.41,"stddev":0.23},"prefillTokensPerSecondTotal":{"mean":6669.33,"stddev":5.37},"decodeTokensPerSecondPerUser":{"mean":5.85,"stddev":1.22}},"sourceRun":"Test 05","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","workloadId":"large-context-25k","sourceLabel":"nvidia/Nemotron-Cascade-2-30B-A3B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"05-big-context","prefixCaching":false},{"id":"nemotron-cascade-2-30b-a3b-bf16-concurrency-stress-25k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":35.919,"stddev":19.476},"decodeTokensPerSecondTotal":{"mean":49.43,"stddev":0.1},"prefillTokensPerSecondTotal":{"mean":6654,"stddev":8.52},"decodeTokensPerSecondPerUser":{"mean":4,"stddev":1.14}},"sourceRun":"Test 06","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","workloadId":"concurrency-stress-25k","sourceLabel":"nvidia/Nemotron-Cascade-2-30B-A3B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"06-concurrency-stress","prefixCaching":false}],"openLoopResults":[{"id":"nemotron-cascade-2-30b-a3b-bf16-office-random-4k","date":"2026-08-09","metrics":{"e2eMs":{"p50":112493.81,"p90":196413.2,"p95":207265.51,"p99":229148.13,"mean":111219.51},"tpotMs":{"p50":231.94,"p90":255.63,"p95":265.06,"p99":282.49,"mean":229.52},"ttftMs":{"p50":1298.29,"p90":2156.32,"p95":2515.31,"p99":2897.81,"mean":1407.29}},"sourceRun":"Test 07","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","scenarioId":"office-random-4k","achievedRps":0.253,"sourceStatus":"complete","sourceTestId":"07-office-baseline","configuredRps":0.3,"prefixCaching":false,"totalRequests":200,"successfulRequests":200,"totalTokenThroughput":1158.61,"peakConcurrentRequests":46},{"id":"nemotron-cascade-2-30b-a3b-bf16-sharegpt-replay","date":"2026-08-09","metrics":{"e2eMs":{"p50":15803.81,"p90":67588.67,"p95":81485.73,"p99":111120.2,"mean":26785.4},"tpotMs":{"p50":134.12,"p90":156.37,"p95":161.31,"p99":185.56,"mean":124.35},"ttftMs":{"p50":489.13,"p90":646.43,"p95":700.4,"p99":743.2,"mean":491.2}},"sourceRun":"Test 08","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","scenarioId":"sharegpt-replay","achievedRps":0.291,"sourceStatus":"complete","sourceTestId":"08-sharegpt","configuredRps":0.3,"prefixCaching":false,"totalRequests":250,"successfulRequests":250,"totalTokenThroughput":133.93,"peakConcurrentRequests":19},{"id":"nemotron-cascade-2-30b-a3b-bf16-reasoning","date":"2026-08-09","metrics":{"e2eMs":{"p50":836339.97,"p90":1186733.5,"p95":1274503.36,"p99":1311854.21,"mean":824209.48},"tpotMs":{"p50":206.17,"p90":234.76,"p95":237.15,"p99":241.71,"mean":204.11},"ttftMs":{"p50":706.45,"p90":921.31,"p95":972.67,"p99":1057.4,"mean":734.19}},"sourceRun":"Test 09","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","scenarioId":"reasoning","achievedRps":0.035,"sourceStatus":"complete","sourceTestId":"09-reasoning","configuredRps":0.2,"prefixCaching":false,"totalRequests":50,"successfulRequests":50,"totalTokenThroughput":188.05,"peakConcurrentRequests":49},{"id":"nemotron-cascade-2-30b-a3b-bf16-monday-peak-random-4k","date":"2026-08-09","metrics":{"e2eMs":{"p50":107370.2,"p90":190074.92,"p95":197509.09,"p99":204455.04,"mean":106909.66},"tpotMs":{"p50":215.17,"p90":220.86,"p95":224.75,"p99":231.84,"mean":212.8},"ttftMs":{"p50":1140.2,"p90":1826.51,"p95":2191.51,"p99":5119.77,"mean":1298.22}},"sourceRun":"Test 10","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","scenarioId":"monday-peak-random-4k","achievedRps":0.225,"sourceStatus":"complete","sourceTestId":"10-monday-peak","configuredRps":1.5,"prefixCaching":false,"totalRequests":300,"successfulRequests":300,"totalTokenThroughput":1019.9,"peakConcurrentRequests":28}],"rateSweep":{"date":"2026-08-13","capacity":{"slo2000":null,"slo5000":{"outputTps":259.3,"ttftP95Ms":4302.7,"achievedRps":0.527,"configuredRps":1,"peakConcurrent":196},"slo10000":{"outputTps":259.3,"ttftP95Ms":4302.7,"achievedRps":0.527,"configuredRps":1,"peakConcurrent":196}},"sourceRun":"Test 11","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/nemotron-cascade-2/nemotron-cascade-2-30b-a3b/bf16/","abortedAtRps":null,"queueKneeRps":null,"sourceStatus":"complete","sourceTestId":"11-rate-sweep","totalRequests":850,"successfulRequests":850}}