Submission Reports
Select a report to view detailed benchmark data, charts, and run metrics.
| Submitter | System Name | Model | Precision | Framework | Division | Runs | Submitted | |
|---|---|---|---|---|---|---|---|---|
| NVIDIA Corporation | Qwen3.6 35B-A3B | vLLM Qwen-latest 7cc4a74c3427 behind Dynamo 1.4.2 | Standardized | 9 | 10/15/2026 | |||
| Google LLC | Qwen3.6 35B-A3B | vLLM Qwen-latest 7cc4a74c3427 behind Dynamo 1.4.2 (4 x TP1 aggregated workers) | Standardized | 9 | 10/16/2026 | |||
| NVIDIA | Qwen3.6 35B-A3B | vLLM:d48687fdeedd | Standardized | 11 | 10/17/2026 | |||
| NVIDIA | GPT-OSS 120B | TensorRT-LLM 1.3.0rc29 (GPT-OSS branch feat/gpt-oss_gb300_26.08) + Dynamo 1.4.2 | Standardized | 11 | 11/13/2026 | |||
| Krai | deepseek-v4_1-flash | vLLM v0.30.0 (ced6857a) | Standardized | 9 | 10/13/2026 | |||
| NVIDIA Corporation | NVIDIA GB300 NVL72 18
Nodes | GPT-OSS 120B | TensorRT-LLM | Standardized | 14 | 06/25/2026 | ||
| NVIDIA Corporation | NVIDIA GB300 NVL72 18
Nodes | DeepSeek-R1 | TensorRT-LLM | Standardized | 14 | 06/25/2026 | ||
| Krai | deepseek-v4_1-flash | vLLM v0.31.1rc1 (nightly build 5555a2d8f; ROCm 10.0.0 stable release) + AITER 0.1.24.post1 with the gfx942 MoE a16w4 kernel patches in shared/src/mi300x_vllm/patches | Standardized | 7 | 11/03/2026 | |||
| Azure | ND MI300X-v5 1
Node | DeepSeek-R1 | vLLM 0.31.0 (image vllm/vllm-openai-rocm@sha256:749f6f3f944f12af49966ac523c1f8573e4b229594b541954bd5f879b4496b1f) | Standardized | 8 | 11/10/2026 | ||
| Google LLC | deepseek-v4_1-flash | vLLM 0.20.2rc1.dev6106+g7cf5542c2.d20261003 behind Dynamo (DEP4: 4 x TP1, DP4, EP4) | Standardized | 9 | 10/16/2026 | |||
| Intel Corporation | 4x Arc(TM) Pro B70 1
Node | Llama 3.1 8B | FP16 OSL: 128, ISL: 870 | vLLM 0.1.dev11435+gfa77e1a63 | Standardized | 9 | 06/26/2026 | |
| NVIDIA Corporation | NVIDIA GB200 NVL72 18
Nodes | DeepSeek-R1 | TensorRT-LLM | Standardized | 11 | 06/25/2026 | ||
| NVIDIA | deepseek-v4_1-flash | mxfp4 | vLLM (zihaok/vllm d63b8107b7081598c098bb34f1acc995f7ff36f5) + Dynamo 1.6.0.dev20261016 | Standardized | 10 | 11/10/2026 | ||
| NVIDIA | kimi-k3 | TensorRT-LLM 1.3.0rc25.post1 + Dynamo 1.6.0.dev20260918 | Standardized | 10 | 11/07/2026 | |||
| Google LLC | DeepSeek-R1 | TensorRT-LLM | Standardized | 9 | 06/26/2026 | |||
| NVIDIA | deepseek-v4_1-flash | vLLM 0.20.2rc1.dev6106+g7cf5542c2 behind Dynamo 1.6.0.dev20260924 | Standardized | 10 | 10/17/2026 | |||
| Krai | deepseek-v4_1-flash | vLLM v0.32.0 (5df63a72) with a custom patch (shared src: src/benchmarking/docker/patches/v0.32.0) | Standardized | 7 | 11/10/2026 | |||
| CoreWeave | NVIDIA GB200 NVL72 17
Nodes | DeepSeek-R1 | nvfp4 | Standardized | 10 | 06/25/2026 | ||
| NVIDIA Corporation | kimi-k3 | TensorRT-LLM 1.3.0rc25.post1 + Dynamo 1.6.0.dev20260918 | Standardized | 13 | 10/16/2026 | |||
| Krai | deepseek-v4_1-flash | vLLM v0.31.1rc1 (nightly build 5555a2d8f; ROCm 10.0.0 stable release) + AITER 0.1.24.post1 with the gfx942 MoE a16w4 kernel patches in shared/src/mi300x_vllm/patches | Standardized | 7 | 10/27/2026 | |||
| Krai | deepseek-v4_1-flash | vLLM v0.31.1rc1 (public ROCm nightly build 5555a2d8f, ROCm 10.0.0) + AITER 0.1.24.post1 as shipped in the image; no patches | Standardized | 7 | 10/27/2026 | |||
| Google LLC | DeepSeek-R1 | TensorRT-LLM 1.3.0rc14 (disaggregated prefill/decode) | Standardized | 10 | 10/16/2026 | |||
| NVIDIA Corporation | kimi-k3 | TensorRT-LLM 1.3.0rc25.post1 + Dynamo 1.6.0.dev20260918 | Standardized | 16 | 10/15/2026 | |||
| Google LLC | kimi-k3 | TensorRT-LLM 1.3.0rc25.post1 + Dynamo 1.6.0.dev20260918 | Standardized | 10 | 10/16/2026 | |||
| Krai Ltd | Dell PowerEdge XE9680 1
Node | Llama 3.1 8B | bfloat16 OSL: 60, ISL: 800 | vLLM v0.19.1 | Standardized | 8 | 06/25/2026 | |
| Krai | deepseek-v4_1-flash | vLLM v0.32.0 (5df63a72) | Standardized | 7 | 11/10/2026 | |||
| NVIDIA Corporation | deepseek-v4_1-flash | vLLM 0.20.2rc1.dev6106+g7cf5542c2.d20261003 behind Dynamo | Standardized | 9 | 10/15/2026 | |||
| Intel Corporation | 4x Arc(TM) Pro B70 1
Node | GPT-OSS 120B | FP16 OSL: 1278, ISL: 5011 | vLLM 0.1.dev11435+gfa77e1a63 | Standardized | 7 | 06/26/2026 | |
| Krai Ltd | Dell PowerEdge XE9680 1
Node | Llama 3.1 8B | bfloat16 OSL: 60, ISL: 800 | vLLM v0.19.1 | Standardized | 8 | 06/26/2026 | |
| Intel Corporation | 1-node-2S-GNR_128C 1
Node | Qwen3.6 35B-A3B | INT8 | vLLM 0.30.1rc1.dev57+g67ddac3ea; vLLM 0.31.1rc1.dev366+g7a0d68f29; vLLM 0.32.0rc1.dev8+g7a0d68f29.d20261018 | Standardized | 7 | 10/21/2026 | |
| Google LLC | GPT-OSS 120B | TensorRT-LLM 1.3.0rc14 (disaggregated prefill/decode) | Standardized | 8 | 10/16/2026 | |||
| Azure | ND MI300X-v5 1
Node | GPT-OSS 120B | vLLM 0.31.0 (image vllm/vllm-openai-rocm@sha256:749f6f3f944f12af49966ac523c1f8573e4b229594b541954bd5f879b4496b1f) | Standardized | 8 | 10/28/2026 | ||
| NVIDIA | Qwen3.6 35B-A3B | vLLM 1af1c649355f behind Dynamo 1.6.0.dev20261021 vLLM sidecars | Standardized | 11 | 11/07/2026 | |||
| Intel Corporation | 1-node-2S-GNR_128C 1
Node | Llama 3.1 8B | FP16 OSL: 128, ISL: 870 | vLLM 0.23.0 | Standardized | 10 | 06/26/2026 |