{"name":"Mistral-Small 4 119B","arena":{"id":"mistral-small-4-119b-nvfp4","name":"Mistral-Small 4 119B","vram":83,"params":119,"vendor":"mistral","quality":{"coding":null,"science":{"bench":"GPQA-Diamond","value":71.2},"knowledge":null},"contextK":256},"notes":"Model-ID: mistralai/Mistral-Small-4-119B-2603-NVFP4. Container: vllm/vllm-openai:v0.26.0-aarch64-cu129-ubuntu2404. llama-benchy 0.4.0. Bronrun: 2026-08-13T10:05:20+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 --tokenizer_mode mistral --config_format mistral --load_format mistral --moe-backend marlin --linear-backend marlin --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":150,"engine":"vLLM v0.26.0","verdict":"Gemeten op 2026-08-13 met vLLM v0.26.0. In de chattest met tien gelijktijdige verzoeken haalt dit profiel 11,62 tokens/s per gebruiker, met gemiddeld 2,14 seconden tot het eerste token. Bij 25k context stijgt die gemiddelde wachttijd naar 38,75 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/mistralai/Mistral-Small-4-119B-2603-NVFP4","provider":"mistralai","precision":"NVFP4","parameters":"119B-A6.5B","provenance":{"modelId":"mistralai/Mistral-Small-4-119B-2603-NVFP4","runPath":"results/mistral-small-4/mistral-small-4-119b/nvfp4","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/mistral-small-4/mistral-small-4-119b/nvfp4/","servedName":"mistral-small-4-119b-nvfp4","generatedAt":"2026-08-13T10:05:20+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":"mistral-small-4-119b-nvfp4-chat-1k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":2.136,"stddev":0.401},"decodeTokensPerSecondTotal":{"mean":80.21,"stddev":2.56},"prefillTokensPerSecondTotal":{"mean":3718.61,"stddev":63.32},"decodeTokensPerSecondPerUser":{"mean":11.62,"stddev":1.35}},"sourceRun":"Test 01","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","workloadId":"chat-1k","sourceLabel":"mistralai/Mistral-Small-4-119B-2603-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"01-chat","prefixCaching":false},{"id":"mistral-small-4-119b-nvfp4-rag-8k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":11.125,"stddev":5.335},"decodeTokensPerSecondTotal":{"mean":60.5,"stddev":3.84},"prefillTokensPerSecondTotal":{"mean":4085.73,"stddev":3.78},"decodeTokensPerSecondPerUser":{"mean":8.99,"stddev":2.26}},"sourceRun":"Test 02","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","workloadId":"rag-8k","sourceLabel":"mistralai/Mistral-Small-4-119B-2603-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"02-rag-8k","prefixCaching":false},{"id":"mistral-small-4-119b-nvfp4-long-output-agents","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":0.697,"stddev":0.012},"decodeTokensPerSecondTotal":{"mean":82.87,"stddev":2.97},"prefillTokensPerSecondTotal":{"mean":3321.26,"stddev":24.66},"decodeTokensPerSecondPerUser":{"mean":12.17,"stddev":1.33}},"sourceRun":"Test 03","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","workloadId":"long-output-agents","sourceLabel":"mistralai/Mistral-Small-4-119B-2603-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"03-long-output","prefixCaching":false},{"id":"mistral-small-4-119b-nvfp4-multi-turn-office","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":3.512,"stddev":0.979},"decodeTokensPerSecondTotal":{"mean":88.73,"stddev":1.12},"prefillTokensPerSecondTotal":{"mean":4103.21,"stddev":11.61},"decodeTokensPerSecondPerUser":{"mean":9.96,"stddev":0.43}},"sourceRun":"Test 04","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","workloadId":"multi-turn-office","sourceLabel":"mistralai/Mistral-Small-4-119B-2603-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"04-multi-turn","prefixCaching":false},{"id":"mistral-small-4-119b-nvfp4-large-context-25k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":38.75,"stddev":19.238},"decodeTokensPerSecondTotal":{"mean":22.14,"stddev":1.66},"prefillTokensPerSecondTotal":{"mean":3303.19,"stddev":4.57},"decodeTokensPerSecondPerUser":{"mean":4.6,"stddev":2.52}},"sourceRun":"Test 05","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","workloadId":"large-context-25k","sourceLabel":"mistralai/Mistral-Small-4-119B-2603-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"05-big-context","prefixCaching":false},{"id":"mistral-small-4-119b-nvfp4-concurrency-stress-25k","date":"2026-08-13","runs":3,"metrics":{"ttfrSec":{"mean":73.818,"stddev":39.602},"decodeTokensPerSecondTotal":{"mean":22.21,"stddev":0.36},"prefillTokensPerSecondTotal":{"mean":3266.33,"stddev":3.11},"decodeTokensPerSecondPerUser":{"mean":2.56,"stddev":1.72}},"sourceRun":"Test 06","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","workloadId":"concurrency-stress-25k","sourceLabel":"mistralai/Mistral-Small-4-119B-2603-NVFP4 nvfp4 op de DGX Spark","sourceStatus":"complete","sourceTestId":"06-concurrency-stress","prefixCaching":false}],"openLoopResults":[{"id":"mistral-small-4-119b-nvfp4-office-random-4k","date":"2026-08-13","metrics":{"e2eMs":{"p50":151856.58,"p90":261804.76,"p95":278171.76,"p99":308844.27,"mean":150442.14},"tpotMs":{"p50":326.75,"p90":353.46,"p95":369.04,"p99":376.25,"mean":311.67},"ttftMs":{"p50":1984.18,"p90":3610.91,"p95":4219.62,"p99":5346.73,"mean":2205.82}},"sourceRun":"Test 07","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","scenarioId":"office-random-4k","achievedRps":0.248,"sourceStatus":"complete","sourceTestId":"07-office-baseline","configuredRps":0.3,"prefixCaching":false,"totalRequests":200,"successfulRequests":200,"totalTokenThroughput":1133.1,"peakConcurrentRequests":55},{"id":"mistral-small-4-119b-nvfp4-sharegpt-replay","date":"2026-08-13","metrics":{"e2eMs":{"p50":8603.75,"p90":39658.39,"p95":47360.99,"p99":64523.24,"mean":15646.26},"tpotMs":{"p50":79.75,"p90":105.25,"p95":113.98,"p99":138.93,"mean":77.29},"ttftMs":{"p50":402.31,"p90":566.33,"p95":624.36,"p99":679.24,"mean":405.13}},"sourceRun":"Test 08","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","scenarioId":"sharegpt-replay","achievedRps":0.292,"sourceStatus":"complete","sourceTestId":"08-sharegpt","configuredRps":0.3,"prefixCaching":false,"totalRequests":250,"successfulRequests":250,"totalTokenThroughput":126.54,"peakConcurrentRequests":13},{"id":"mistral-small-4-119b-nvfp4-reasoning","date":"2026-08-13","metrics":{"e2eMs":{"p50":963871.99,"p90":1364522.08,"p95":1442413.28,"p99":1478563.46,"mean":941921.59},"tpotMs":{"p50":231.56,"p90":263.21,"p95":265.04,"p99":270.48,"mean":230.92},"ttftMs":{"p50":763.22,"p90":1057.83,"p95":1069.06,"p99":1158.85,"mean":775.14}},"sourceRun":"Test 09","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","scenarioId":"reasoning","achievedRps":0.032,"sourceStatus":"complete","sourceTestId":"09-reasoning","configuredRps":0.2,"prefixCaching":false,"totalRequests":50,"successfulRequests":50,"totalTokenThroughput":167.76,"peakConcurrentRequests":49},{"id":"mistral-small-4-119b-nvfp4-monday-peak-random-4k","date":"2026-08-13","metrics":{"e2eMs":{"p50":113446.6,"p90":203410.78,"p95":211714.92,"p99":222903.96,"mean":114787.71},"tpotMs":{"p50":229.69,"p90":239.73,"p95":249.07,"p99":272.49,"mean":227.94},"ttftMs":{"p50":1548.51,"p90":2493.41,"p95":5651.16,"p99":12199.84,"mean":2032.58}},"sourceRun":"Test 10","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","scenarioId":"monday-peak-random-4k","achievedRps":0.211,"sourceStatus":"complete","sourceTestId":"10-monday-peak","configuredRps":1.5,"prefixCaching":false,"totalRequests":300,"successfulRequests":300,"totalTokenThroughput":951.3,"peakConcurrentRequests":27}],"rateSweep":{"date":"2026-08-13","capacity":{"slo2000":null,"slo5000":{"outputTps":107.8,"ttftP95Ms":4748.7,"achievedRps":0.219,"configuredRps":0.3,"peakConcurrent":46},"slo10000":{"outputTps":162,"ttftP95Ms":7717.5,"achievedRps":0.332,"configuredRps":0.7,"peakConcurrent":156}},"sourceRun":"Test 11","sourceUrl":"https://github.com/djangodevreng/dgx-spark-benchmarks/tree/a948cec23b054c62fef595bad3cdb284066922ac/results/mistral-small-4/mistral-small-4-119b/nvfp4/","abortedAtRps":null,"queueKneeRps":null,"sourceStatus":"complete","sourceTestId":"11-rate-sweep","totalRequests":850,"successfulRequests":850}}