{"name":"Qwen-3.6-27B","arena":{"id":"qwen-3-6-27b-bf16","name":"Qwen-3.6 27B","vram":54,"params":27,"vendor":"alibaba","quality":{"coding":{"bench":"LiveCodeBench v6","value":83.9},"science":{"bench":"GPQA-Diamond","value":87.8},"knowledge":{"bench":"MMLU-Pro","value":86.2}},"contextK":256,"baseModelId":"qwen-3-6-27b"},"notes":"Model-ID: Qwen/Qwen3.6-27B. Container: vllm/vllm-openai:v0.26.0-aarch64-cu129-ubuntu2404. llama-benchy 0.4.0. Bronrun: 2026-08-13T19:58:59+02:00. KV-cache fp8; max_model_len 131072; gpu_memory_utilization 0,9. Prefix caching uit; async scheduling aan. Extra serverflags: --language-model-only --reasoning-parser qwen3 --trust-remote-code. 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":320,"engine":"vLLM v0.26.0","verdict":"Gemeten op 2026-08-13 tot 2026-08-14 met vLLM v0.26.0. In de chattest met tien gelijktijdige verzoeken haalt dit profiel 4,06 tokens/s per gebruiker, met gemiddeld 5,88 seconden tot het eerste token. Bij 25k context stijgt die gemiddelde wachttijd naar 120,4 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/Qwen/Qwen3.6-27B-FP8","provider":"Alibaba","precision":"BF16","parameters":"27B","provenance":{"modelId":"Qwen/Qwen3.6-27B","runPath":"results/qwen-3.6/qwen-3.6-27b/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/qwen-3.6/qwen-3.6-27b/bf16/","servedName":"qwen3.6-27b-bf16","generatedAt":"2026-08-13T19:58:59+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":"Hybrid","measuredContextK":128,"benchmarkResults":[{"id":"qwen-3-6-27b-bf16-chat-1k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":5.883,"stddev":2.248},"decodeTokensPerSecondTotal":{"mean":39.82,"stddev":0.01},"prefillTokensPerSecondTotal":{"mean":1093.72,"stddev":1.05},"decodeTokensPerSecondPerUser":{"mean":4.06,"stddev":0.03}},"sourceRun":"Test 01","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","workloadId":"chat-1k","sourceLabel":"Qwen/Qwen3.6-27B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"01-chat","prefixCaching":false},{"id":"qwen-3-6-27b-bf16-rag-8k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":37.902,"stddev":18.916},"decodeTokensPerSecondTotal":{"mean":27.25,"stddev":0.08},"prefillTokensPerSecondTotal":{"mean":1110.87,"stddev":0.99},"decodeTokensPerSecondPerUser":{"mean":3.36,"stddev":0.36}},"sourceRun":"Test 02","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","workloadId":"rag-8k","sourceLabel":"Qwen/Qwen3.6-27B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"02-rag-8k","prefixCaching":false},{"id":"qwen-3-6-27b-bf16-long-output-agents","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":2.112,"stddev":0.341},"decodeTokensPerSecondTotal":{"mean":30.92,"stddev":2.98},"prefillTokensPerSecondTotal":{"mean":1001.61,"stddev":36.65},"decodeTokensPerSecondPerUser":{"mean":4.13,"stddev":0.03}},"sourceRun":"Test 03","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","workloadId":"long-output-agents","sourceLabel":"Qwen/Qwen3.6-27B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"03-long-output","prefixCaching":false},{"id":"qwen-3-6-27b-bf16-multi-turn-office","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":10.386,"stddev":4.81},"decodeTokensPerSecondTotal":{"mean":35.72,"stddev":1.23},"prefillTokensPerSecondTotal":{"mean":1122.73,"stddev":2.77},"decodeTokensPerSecondPerUser":{"mean":3.92,"stddev":0.12}},"sourceRun":"Test 04","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","workloadId":"multi-turn-office","sourceLabel":"Qwen/Qwen3.6-27B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"04-multi-turn","prefixCaching":false},{"id":"qwen-3-6-27b-bf16-large-context-25k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":120.398,"stddev":61.971},"decodeTokensPerSecondTotal":{"mean":9.75,"stddev":0.03},"prefillTokensPerSecondTotal":{"mean":1043.61,"stddev":1.17},"decodeTokensPerSecondPerUser":{"mean":1.95,"stddev":0.83}},"sourceRun":"Test 05","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","workloadId":"large-context-25k","sourceLabel":"Qwen/Qwen3.6-27B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"05-big-context","prefixCaching":false},{"id":"qwen-3-6-27b-bf16-concurrency-stress-25k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":231.009,"stddev":127.007},"decodeTokensPerSecondTotal":{"mean":10.33,"stddev":0.01},"prefillTokensPerSecondTotal":{"mean":1028.82,"stddev":0.5},"decodeTokensPerSecondPerUser":{"mean":1.26,"stddev":0.73}},"sourceRun":"Test 06","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","workloadId":"concurrency-stress-25k","sourceLabel":"Qwen/Qwen3.6-27B bf16 op de DGX Spark","sourceStatus":"complete","sourceTestId":"06-concurrency-stress","prefixCaching":false}],"openLoopResults":[{"id":"qwen-3-6-27b-bf16-office-random-4k","date":"2026-08-14","metrics":{"e2eMs":{"p50":619407,"p90":914721.17,"p95":983166.02,"p99":1047452.89,"mean":617303.9},"tpotMs":{"p50":1258.14,"p90":2001.32,"p95":2039.54,"p99":2094.06,"mean":1280.79},"ttftMs":{"p50":82171.42,"p90":141781.37,"p95":149996.11,"p99":169012.02,"mean":81481.57}},"sourceRun":"Test 07","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","scenarioId":"office-random-4k","achievedRps":0.164,"sourceStatus":"complete","sourceTestId":"07-office-baseline","configuredRps":0.3,"prefixCaching":false,"totalRequests":200,"successfulRequests":200,"totalTokenThroughput":747.55,"peakConcurrentRequests":175},{"id":"qwen-3-6-27b-bf16-sharegpt-replay","date":"2026-08-14","metrics":{"e2eMs":{"p50":44099.86,"p90":170232.21,"p95":218517.85,"p99":279929.42,"mean":69871.65},"tpotMs":{"p50":338.13,"p90":352.29,"p95":360.89,"p99":389.09,"mean":326.45},"ttftMs":{"p50":988.98,"p90":1394.91,"p95":1510.46,"p99":1649.74,"mean":1026.13}},"sourceRun":"Test 08","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","scenarioId":"sharegpt-replay","achievedRps":0.261,"sourceStatus":"complete","sourceTestId":"08-sharegpt","configuredRps":0.3,"prefixCaching":false,"totalRequests":250,"successfulRequests":250,"totalTokenThroughput":115.15,"peakConcurrentRequests":34},{"id":"qwen-3-6-27b-bf16-reasoning","date":"2026-08-14","metrics":{"e2eMs":{"p50":1455182.43,"p90":2259298.97,"p95":2428817.89,"p99":2482488.39,"mean":1503170.21},"tpotMs":{"p50":371.09,"p90":381.44,"p95":385.9,"p99":390.82,"mean":365.54},"ttftMs":{"p50":1777,"p90":2587.08,"p95":2865.71,"p99":3200.72,"mean":1873.13}},"sourceRun":"Test 09","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","scenarioId":"reasoning","achievedRps":0.019,"sourceStatus":"complete","sourceTestId":"09-reasoning","configuredRps":0.2,"prefixCaching":false,"totalRequests":50,"successfulRequests":50,"totalTokenThroughput":101.42,"peakConcurrentRequests":50},{"id":"qwen-3-6-27b-bf16-monday-peak-random-4k","date":"2026-08-14","metrics":{"e2eMs":{"p50":239206.79,"p90":434741.53,"p95":452052.42,"p99":484486.06,"mean":246903.97},"tpotMs":{"p50":483.98,"p90":523.77,"p95":552.2,"p99":639.53,"mean":483.38},"ttftMs":{"p50":5421.85,"p90":10147.18,"p95":28294.76,"p99":74364.67,"mean":8342.2}},"sourceRun":"Test 10","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","scenarioId":"monday-peak-random-4k","achievedRps":0.098,"sourceStatus":"complete","sourceTestId":"10-monday-peak","configuredRps":1.5,"prefixCaching":false,"totalRequests":300,"successfulRequests":300,"totalTokenThroughput":440.89,"peakConcurrentRequests":28}],"rateSweep":{"date":"2026-08-13","capacity":{"slo2000":null,"slo5000":null,"slo10000":null},"sourceRun":"Test 11","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/qwen-3.6/qwen-3.6-27b/bf16/","abortedAtRps":null,"queueKneeRps":null,"sourceStatus":"complete","sourceTestId":"11-rate-sweep","totalRequests":850,"successfulRequests":850}}