{"name":"Muse-Glimmer 30B","arena":{"id":"muse-glimmer-30b-bf16-spec","name":"Muse-Glimmer 30B","vram":60,"params":30,"vendor":"meta","quality":{"coding":null,"science":{"bench":"GPQA-Diamond","value":83.5},"knowledge":null},"contextK":128,"baseModelId":"muse-glimmer-30b"},"notes":"Model-ID: meta-models/Muse-Glimmer-30B. Container: vllm/vllm-openai:muse-glimmer. llama-benchy 0.4.0. Bronrun: 2026-08-12T14:16:38+02:00. KV-cache auto; max_model_len 131072; gpu_memory_utilization 0,9. Prefix caching uit; async scheduling uit. Extra serverflags: --max-num-batched-tokens 8192 --max-num-seqs 256 --generation-config auto --enable-auto-tool-choice --tool-call-parser muse_glimmer --reasoning-parser muse_glimmer --speculative-config {\"method\":\"dflash\",\"model\":\"meta-models/Muse-Glimmer-30B-assistant\",\"num_speculative_tokens\":15}. 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":170,"engine":"vLLM v0.26.1rc1.dev608+g99a10304d","verdict":"Gemeten op 2026-08-12 met vLLM v0.26.1rc1.dev608+g99a10304d. In de chattest met tien gelijktijdige verzoeken haalt dit profiel 13,3 tokens/s per gebruiker, met gemiddeld 5,67 seconden tot het eerste token. Bij 25k context stijgt die gemiddelde wachttijd naar 89,78 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/meta-models/Muse-Glimmer-30B","provider":"meta-models","precision":"BF16","parameters":"30B","provenance":{"modelId":"meta-models/Muse-Glimmer-30B","runPath":"results/muse-glimmer/muse-glimmer-30b/bf16-spec","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/muse-glimmer/muse-glimmer-30b/bf16-spec/","servedName":"muse-glimmer-30b-bf16-spec","generatedAt":"2026-08-12T14:16:38+02:00","runRevision":"a948cec23b054c62fef595bad3cdb284066922ac","vllmVersion":"0.26.1rc1.dev608+g99a10304d","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":"Dense","measuredContextK":128,"benchmarkResults":[{"id":"muse-glimmer-30b-bf16-spec-chat-1k","date":"2026-08-12","runs":3,"metrics":{"ttfrSec":{"mean":5.667,"stddev":1.017},"decodeTokensPerSecondTotal":{"mean":86.23,"stddev":5.37},"prefillTokensPerSecondTotal":{"mean":1349.33,"stddev":16.5},"decodeTokensPerSecondPerUser":{"mean":13.3,"stddev":6.78}},"sourceRun":"Test 01","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","workloadId":"chat-1k","sourceLabel":"meta-models/Muse-Glimmer-30B bf16-spec op de DGX Spark","sourceStatus":"complete","sourceTestId":"01-chat","prefixCaching":false},{"id":"muse-glimmer-30b-bf16-spec-rag-8k","date":"2026-08-12","runs":3,"metrics":{"ttfrSec":{"mean":29.45,"stddev":14.994},"decodeTokensPerSecondTotal":{"mean":43.99,"stddev":1.32},"prefillTokensPerSecondTotal":{"mean":1458.94,"stddev":7.5},"decodeTokensPerSecondPerUser":{"mean":6.63,"stddev":1.42}},"sourceRun":"Test 02","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","workloadId":"rag-8k","sourceLabel":"meta-models/Muse-Glimmer-30B bf16-spec op de DGX Spark","sourceStatus":"complete","sourceTestId":"02-rag-8k","prefixCaching":false},{"id":"muse-glimmer-30b-bf16-spec-long-output-agents","date":"2026-08-12","runs":3,"metrics":{"ttfrSec":{"mean":1.896,"stddev":0.015},"decodeTokensPerSecondTotal":{"mean":89.62,"stddev":4.78},"prefillTokensPerSecondTotal":{"mean":966.87,"stddev":36.43},"decodeTokensPerSecondPerUser":{"mean":12.07,"stddev":1.42}},"sourceRun":"Test 03","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","workloadId":"long-output-agents","sourceLabel":"meta-models/Muse-Glimmer-30B bf16-spec op de DGX Spark","sourceStatus":"complete","sourceTestId":"03-long-output","prefixCaching":false},{"id":"muse-glimmer-30b-bf16-spec-multi-turn-office","date":"2026-08-12","runs":3,"metrics":{"ttfrSec":{"mean":9.079,"stddev":3.372},"decodeTokensPerSecondTotal":{"mean":63.92,"stddev":2.58},"prefillTokensPerSecondTotal":{"mean":1420.93,"stddev":9.07},"decodeTokensPerSecondPerUser":{"mean":13.19,"stddev":9.16}},"sourceRun":"Test 04","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","workloadId":"multi-turn-office","sourceLabel":"meta-models/Muse-Glimmer-30B bf16-spec op de DGX Spark","sourceStatus":"complete","sourceTestId":"04-multi-turn","prefixCaching":false},{"id":"muse-glimmer-30b-bf16-spec-large-context-25k","date":"2026-08-12","runs":3,"metrics":{"ttfrSec":{"mean":89.777,"stddev":45.655},"decodeTokensPerSecondTotal":{"mean":14.45,"stddev":0.1},"prefillTokensPerSecondTotal":{"mean":1411.69,"stddev":1.7},"decodeTokensPerSecondPerUser":{"mean":3.38,"stddev":1.94}},"sourceRun":"Test 05","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","workloadId":"large-context-25k","sourceLabel":"meta-models/Muse-Glimmer-30B bf16-spec op de DGX Spark","sourceStatus":"complete","sourceTestId":"05-big-context","prefixCaching":false},{"id":"muse-glimmer-30b-bf16-spec-concurrency-stress-25k","date":"2026-08-12","runs":3,"metrics":{"ttfrSec":{"mean":171.624,"stddev":94.35},"decodeTokensPerSecondTotal":{"mean":14.13,"stddev":0.05},"prefillTokensPerSecondTotal":{"mean":1372.42,"stddev":2.15},"decodeTokensPerSecondPerUser":{"mean":1.82,"stddev":1.23}},"sourceRun":"Test 06","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","workloadId":"concurrency-stress-25k","sourceLabel":"meta-models/Muse-Glimmer-30B bf16-spec op de DGX Spark","sourceStatus":"complete","sourceTestId":"06-concurrency-stress","prefixCaching":false}],"openLoopResults":[{"id":"muse-glimmer-30b-bf16-spec-office-random-4k","date":"2026-08-12","metrics":{"e2eMs":{"p50":1445002.79,"p90":2043613.64,"p95":2132778.08,"p99":2337288.23,"mean":1390507.5},"tpotMs":{"p50":3024.47,"p90":4852.51,"p95":5129.21,"p99":5357.15,"mean":3080.16},"ttftMs":{"p50":59120.38,"p90":221463.51,"p95":263838.8,"p99":316295.52,"mean":85955.93}},"sourceRun":"Test 07","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","scenarioId":"office-random-4k","achievedRps":0.081,"sourceStatus":"complete","sourceTestId":"07-office-baseline","configuredRps":0.3,"prefixCaching":false,"totalRequests":200,"successfulRequests":200,"totalTokenThroughput":373.47,"peakConcurrentRequests":190},{"id":"muse-glimmer-30b-bf16-spec-sharegpt-replay","date":"2026-08-12","metrics":{"e2eMs":{"p50":10747.45,"p90":55921.41,"p95":68395.69,"p99":86652.44,"mean":20554.27},"tpotMs":{"p50":77.4,"p90":129.64,"p95":138.58,"p99":149.14,"mean":77.25},"ttftMs":{"p50":940.67,"p90":1401.63,"p95":1558.71,"p99":2721.09,"mean":1044.77}},"sourceRun":"Test 08","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","scenarioId":"sharegpt-replay","achievedRps":0.286,"sourceStatus":"complete","sourceTestId":"08-sharegpt","configuredRps":0.3,"prefixCaching":false,"totalRequests":250,"successfulRequests":250,"totalTokenThroughput":133.95,"peakConcurrentRequests":13},{"id":"muse-glimmer-30b-bf16-spec-reasoning","date":"2026-08-12","metrics":{"e2eMs":{"p50":922910.53,"p90":1203974.54,"p95":1391017.47,"p99":1572005.88,"mean":904739.04},"tpotMs":{"p50":230.54,"p90":386.5,"p95":423.03,"p99":488.73,"mean":248.51},"ttftMs":{"p50":2601.98,"p90":4153.71,"p95":4338.38,"p99":4473.18,"mean":2727.65}},"sourceRun":"Test 09","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","scenarioId":"reasoning","achievedRps":0.029,"sourceStatus":"complete","sourceTestId":"09-reasoning","configuredRps":0.2,"prefixCaching":false,"totalRequests":50,"successfulRequests":50,"totalTokenThroughput":154.75,"peakConcurrentRequests":49},{"id":"muse-glimmer-30b-bf16-spec-monday-peak-random-4k","date":"2026-08-12","metrics":{"e2eMs":{"p50":253171.91,"p90":619580.8,"p95":672133.6,"p99":730068.8,"mean":300383.29},"tpotMs":{"p50":710.23,"p90":781.44,"p95":796.06,"p99":910.54,"mean":597.52},"ttftMs":{"p50":4857.4,"p90":8468.75,"p95":28022.2,"p99":59273.13,"mean":7343.04}},"sourceRun":"Test 10","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","scenarioId":"monday-peak-random-4k","achievedRps":0.08,"sourceStatus":"complete","sourceTestId":"10-monday-peak","configuredRps":1.5,"prefixCaching":false,"totalRequests":300,"successfulRequests":300,"totalTokenThroughput":362.96,"peakConcurrentRequests":27}],"rateSweep":{"date":"2026-08-12","capacity":{"slo2000":null,"slo5000":null,"slo10000":null},"sourceRun":"Test 11","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/muse-glimmer/muse-glimmer-30b/bf16-spec/","abortedAtRps":null,"queueKneeRps":0.1,"sourceStatus":"complete","sourceTestId":"11-rate-sweep","totalRequests":850,"successfulRequests":850}}