item_id stringclasses 40
values | sample_index int64 0 7 | repo stringclasses 2
values | agent_name stringclasses 2
values | agent_model stringclasses 2
values | human_commit stringclasses 40
values | parent_commit stringclasses 40
values | benchmark_mode stringclasses 3
values | perf_command stringclasses 24
values | llm_model stringclasses 15
values | status stringclasses 1
value | duration_s float64 19.2 418 | ttft_mean_ms float64 38.7 1.13k ⌀ | ttft_median_ms float64 34.5 1.18k ⌀ | ttft_p99_ms float64 106 1.83k ⌀ | tpot_mean_ms float64 6.45 50.6 ⌀ | tpot_median_ms float64 6.42 54 ⌀ | tpot_p99_ms float64 7.18 315 ⌀ | itl_mean_ms float64 6.45 78.3 ⌀ | itl_median_ms float64 5.62 44 ⌀ | itl_p99_ms float64 14.4 1.22k ⌀ | request_throughput_req_s float64 4.87 181 ⌀ | output_token_throughput_tok_s float64 1.04k 11.6k ⌀ | total_token_throughput_tok_s float64 9.39k 57.8k ⌀ | latency_avg_ms float64 215 2.39k ⌀ | latency_p50_ms null | latency_p99_ms null | throughput_tok_s float64 102 9.33k ⌀ | elapsed_time_s float64 3.72 3.81 ⌀ | input_throughput_tok_s float64 17.2k 17.6k ⌀ | timestamp stringdate 2026-03-27 14:05:13 2026-03-31 05:19:27 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
sglang_core-0000 | 0 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 184.103882 | 43.49 | 35.19 | 281.89 | null | null | null | 12.62 | 11.23 | 36.34 | 7.02 | 1,274.12 | null | null | null | null | null | null | null | 2026-03-30T14:52:45.312223+00:00 |
sglang_core-0000 | 1 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 121.800966 | 39.29 | 35.22 | 109.95 | null | null | null | 12.61 | 11.23 | 36.19 | 7.02 | 1,273.97 | null | null | null | null | null | null | null | 2026-03-30T14:54:47.113997+00:00 |
sglang_core-0000 | 2 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 121.476817 | 39.17 | 34.83 | 117.77 | null | null | null | 12.61 | 11.23 | 36.27 | 7.02 | 1,274.49 | null | null | null | null | null | null | null | 2026-03-30T14:56:48.591596+00:00 |
sglang_core-0000 | 3 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 116.222275 | 38.85 | 34.85 | 110.39 | null | null | null | 12.61 | 11.22 | 36.58 | 7.02 | 1,274.26 | null | null | null | null | null | null | null | 2026-03-30T14:58:44.814523+00:00 |
sglang_core-0000 | 4 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 115.634893 | 39.09 | 34.51 | 110.16 | null | null | null | 12.6 | 11.22 | 36.25 | 7.02 | 1,274.4 | null | null | null | null | null | null | null | 2026-03-30T15:00:40.450066+00:00 |
sglang_core-0000 | 5 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 116.497592 | 38.93 | 34.48 | 115.08 | null | null | null | 12.6 | 11.23 | 36.62 | 7.02 | 1,274.63 | null | null | null | null | null | null | null | 2026-03-30T15:02:36.948330+00:00 |
sglang_core-0000 | 6 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 417.82995 | 39.04 | 34.76 | 106.65 | null | null | null | 12.59 | 11.22 | 36.13 | 7.02 | 1,274.5 | null | null | null | null | null | null | null | 2026-03-30T15:09:34.778871+00:00 |
sglang_core-0000 | 7 | sglang | claude_code | sonnet-4.5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 117.93314 | 39.26 | 34.94 | 112.82 | null | null | null | 12.61 | 11.22 | 36.6 | 7.02 | 1,274.11 | null | null | null | null | null | null | null | 2026-03-30T15:11:32.712577+00:00 |
sglang_core-0000 | 0 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 121.030936 | 43.96 | 34.85 | 301.52 | null | null | null | 12.61 | 11.22 | 36.78 | 7.02 | 1,274.53 | null | null | null | null | null | null | null | 2026-03-30T15:29:33.596996+00:00 |
sglang_core-0000 | 1 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 117.104792 | 39.38 | 35.1 | 111.47 | null | null | null | 12.6 | 11.22 | 35.49 | 7.02 | 1,274.2 | null | null | null | null | null | null | null | 2026-03-30T15:31:30.702368+00:00 |
sglang_core-0000 | 2 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 117.014424 | 39.25 | 34.91 | 108.52 | null | null | null | 12.6 | 11.22 | 36.66 | 7.02 | 1,274.32 | null | null | null | null | null | null | null | 2026-03-30T15:33:27.717418+00:00 |
sglang_core-0000 | 3 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 118.100771 | 38.71 | 34.73 | 112.26 | null | null | null | 12.6 | 11.22 | 36.37 | 7.02 | 1,274.56 | null | null | null | null | null | null | null | 2026-03-30T15:35:25.818807+00:00 |
sglang_core-0000 | 4 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 116.464553 | 38.92 | 34.53 | 112.01 | null | null | null | 12.6 | 11.22 | 35.94 | 7.02 | 1,274.22 | null | null | null | null | null | null | null | 2026-03-30T15:37:22.283934+00:00 |
sglang_core-0000 | 5 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 116.35513 | 39.3 | 35.06 | 115.31 | null | null | null | 12.62 | 11.23 | 36.8 | 7.02 | 1,274.36 | null | null | null | null | null | null | null | 2026-03-30T15:39:18.639674+00:00 |
sglang_core-0000 | 6 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 122.275946 | 38.75 | 34.63 | 117.25 | null | null | null | 12.63 | 11.23 | 36.46 | 7.02 | 1,274.15 | null | null | null | null | null | null | null | 2026-03-30T15:41:20.916211+00:00 |
sglang_core-0000 | 7 | sglang | codex_cli | gpt-5 | 021f76e4f49861b2e9ea9ccff06a46d577e3c548 | 777688b8929c877e4e28c2eac208d776abe4c3af | serving | python3 -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompt 480 --request-rate 8 | meta-llama/Llama-3.1-8B-Instruct | success | 117.544699 | 39 | 34.74 | 106.43 | null | null | null | 12.6 | 11.22 | 36.8 | 7.02 | 1,274.54 | null | null | null | null | null | null | null | 2026-03-30T15:43:18.461554+00:00 |
sglang_core-0003 | 0 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 71.840599 | 1,068.55 | 1,064.06 | 1,643.73 | null | null | null | 15.05 | 12.26 | 15.22 | 5.18 | 1,101.36 | null | null | null | null | null | null | null | 2026-03-30T09:54:18.300987+00:00 |
sglang_core-0003 | 1 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 70.986401 | 784.9 | 657.42 | 1,389.94 | null | null | null | 14.89 | 12.25 | 15.26 | 5.29 | 1,124.82 | null | null | null | null | null | null | null | 2026-03-30T09:55:29.287931+00:00 |
sglang_core-0003 | 2 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 71.733032 | 869.73 | 746.8 | 1,414.36 | null | null | null | 14.62 | 12.24 | 16.56 | 5.28 | 1,122.81 | null | null | null | null | null | null | null | 2026-03-30T09:56:41.021460+00:00 |
sglang_core-0003 | 3 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.46572 | 779.3 | 653.41 | 1,374.28 | null | null | null | 14.87 | 12.23 | 15.34 | 5.29 | 1,124.89 | null | null | null | null | null | null | null | 2026-03-30T09:57:47.487710+00:00 |
sglang_core-0003 | 4 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.735435 | 759.21 | 708.45 | 1,313.13 | null | null | null | 14.68 | 12.25 | 16.1 | 5.31 | 1,128.99 | null | null | null | null | null | null | null | 2026-03-30T09:58:53.223650+00:00 |
sglang_core-0003 | 5 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.675062 | 759.44 | 606.15 | 1,340.49 | null | null | null | 14.8 | 12.25 | 15.32 | 5.3 | 1,126.84 | null | null | null | null | null | null | null | 2026-03-30T10:00:00.899319+00:00 |
sglang_core-0003 | 6 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.229535 | 755.06 | 703.63 | 1,310.31 | null | null | null | 14.66 | 12.25 | 15.25 | 5.31 | 1,129.65 | null | null | null | null | null | null | null | 2026-03-30T10:01:07.129548+00:00 |
sglang_core-0003 | 7 | sglang | claude_code | sonnet-4.5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.922524 | 767.55 | 638.79 | 1,370.81 | null | null | null | 14.89 | 12.26 | 15.11 | 5.29 | 1,125.49 | null | null | null | null | null | null | null | 2026-03-30T10:02:13.052692+00:00 |
sglang_core-0003 | 0 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 69.888075 | 780.1 | 711.4 | 1,434.48 | null | null | null | 15.13 | 12.23 | 21.45 | 5.28 | 1,122.21 | null | null | null | null | null | null | null | 2026-03-30T15:53:45.227976+00:00 |
sglang_core-0003 | 1 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 64.586208 | 784.21 | 718.28 | 1,330.74 | null | null | null | 14.73 | 12.25 | 15.59 | 5.28 | 1,122.09 | null | null | null | null | null | null | null | 2026-03-30T15:54:49.814707+00:00 |
sglang_core-0003 | 2 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.577192 | 775.86 | 710.62 | 1,319.87 | null | null | null | 14.69 | 12.24 | 15.22 | 5.29 | 1,124.25 | null | null | null | null | null | null | null | 2026-03-30T15:55:55.392487+00:00 |
sglang_core-0003 | 3 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.87677 | 773.47 | 710.83 | 1,313.85 | null | null | null | 14.68 | 12.26 | 15.54 | 5.29 | 1,124.81 | null | null | null | null | null | null | null | 2026-03-30T15:57:01.269856+00:00 |
sglang_core-0003 | 4 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.298402 | 770.76 | 704.74 | 1,313.94 | null | null | null | 14.67 | 12.24 | 15.52 | 5.29 | 1,124.71 | null | null | null | null | null | null | null | 2026-03-30T15:58:06.568835+00:00 |
sglang_core-0003 | 5 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 64.913679 | 774.24 | 708.63 | 1,316.58 | null | null | null | 14.68 | 12.26 | 15.6 | 5.29 | 1,124.12 | null | null | null | null | null | null | null | 2026-03-30T15:59:11.483068+00:00 |
sglang_core-0003 | 6 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.527048 | 766.82 | 700.55 | 1,312.69 | null | null | null | 14.68 | 12.26 | 15.51 | 5.29 | 1,125.07 | null | null | null | null | null | null | null | 2026-03-30T16:00:17.010729+00:00 |
sglang_core-0003 | 7 | sglang | codex_cli | gpt-5 | 132dad874d2e44592d03a112e4b7d63b153e8346 | 60fdad7cf343333e956a3889c12956396a1516bf | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.095869 | 774.25 | 603.79 | 1,334.1 | null | null | null | 14.76 | 12.25 | 15.95 | 5.28 | 1,122.94 | null | null | null | null | null | null | null | 2026-03-30T16:01:22.107171+00:00 |
sglang_core-0006 | 0 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 102.111309 | 869.89 | 759.08 | 1,393.31 | null | null | null | 15.05 | 12.35 | 16.96 | 4.95 | 1,052.49 | null | null | null | null | null | null | null | 2026-03-30T10:26:35.433323+00:00 |
sglang_core-0006 | 1 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.687716 | 874.18 | 758.05 | 1,424.6 | null | null | null | 15.16 | 12.39 | 16.67 | 4.95 | 1,052.14 | null | null | null | null | null | null | null | 2026-03-30T10:27:41.121876+00:00 |
sglang_core-0006 | 2 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.739752 | 1,122.37 | 1,057.86 | 1,711.2 | null | null | null | 15.37 | 12.36 | 17.19 | 4.87 | 1,035.52 | null | null | null | null | null | null | null | 2026-03-30T10:28:47.862232+00:00 |
sglang_core-0006 | 3 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.573107 | 915.85 | 835.45 | 1,483.29 | null | null | null | 15.28 | 12.33 | 17.74 | 4.91 | 1,044.26 | null | null | null | null | null | null | null | 2026-03-30T10:29:53.435798+00:00 |
sglang_core-0006 | 4 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.609401 | 901.99 | 863.31 | 1,413.11 | null | null | null | 15.08 | 12.36 | 17.09 | 4.94 | 1,050 | null | null | null | null | null | null | null | 2026-03-30T10:30:59.045690+00:00 |
sglang_core-0006 | 5 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.025922 | 894.95 | 805.84 | 1,471.21 | null | null | null | 15.27 | 12.37 | 16.73 | 4.94 | 1,049.81 | null | null | null | null | null | null | null | 2026-03-30T10:32:06.072112+00:00 |
sglang_core-0006 | 6 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.288387 | 884.94 | 765.91 | 1,457.93 | null | null | null | 15.29 | 12.39 | 16.87 | 4.94 | 1,049.58 | null | null | null | null | null | null | null | 2026-03-30T10:33:13.361051+00:00 |
sglang_core-0006 | 7 | sglang | claude_code | sonnet-4.5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.067697 | 886.63 | 896.93 | 1,408.58 | null | null | null | 15.1 | 12.36 | 16.77 | 4.94 | 1,051.42 | null | null | null | null | null | null | null | 2026-03-30T10:34:19.429257+00:00 |
sglang_core-0006 | 0 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 108.676095 | 887.11 | 802.78 | 1,393.22 | null | null | null | 14.97 | 12.39 | 18.4 | 4.94 | 1,050.73 | null | null | null | null | null | null | null | 2026-03-30T16:16:59.764868+00:00 |
sglang_core-0006 | 1 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.147596 | 986.76 | 906.17 | 1,497.87 | null | null | null | 14.98 | 12.36 | 18 | 4.94 | 1,050.3 | null | null | null | null | null | null | null | 2026-03-30T16:18:04.913039+00:00 |
sglang_core-0006 | 2 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 70.658699 | 980.51 | 897.98 | 1,498.82 | null | null | null | 14.99 | 12.37 | 16.98 | 4.93 | 1,049.26 | null | null | null | null | null | null | null | 2026-03-30T16:09:23.447385+00:00 |
sglang_core-0006 | 3 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.239623 | 877.72 | 746.78 | 1,408.78 | null | null | null | 15.1 | 12.38 | 16.62 | 4.95 | 1,052.46 | null | null | null | null | null | null | null | 2026-03-30T16:19:10.153748+00:00 |
sglang_core-0006 | 4 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 70.427929 | 906.02 | 810.12 | 1,434.51 | null | null | null | 15.16 | 12.34 | 17.39 | 4.93 | 1,049.03 | null | null | null | null | null | null | null | 2026-03-30T16:20:20.583285+00:00 |
sglang_core-0006 | 5 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 70.478483 | 1,129.43 | 1,046.02 | 1,707.29 | null | null | null | 15.31 | 12.36 | 16.8 | 4.88 | 1,036.71 | null | null | null | null | null | null | null | 2026-03-30T16:21:31.062341+00:00 |
sglang_core-0006 | 6 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.726401 | 875.01 | 758.8 | 1,431.59 | null | null | null | 15.17 | 12.37 | 16.77 | 4.94 | 1,051.26 | null | null | null | null | null | null | null | 2026-03-30T16:22:36.789354+00:00 |
sglang_core-0006 | 7 | sglang | codex_cli | gpt-5 | 1acca3a2c685221cdb181c2abda4f635e1ead435 | 6ea1e6ac6e2fa949cebd1b4338f9bfb7036d14fe | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.861217 | 870.34 | 753.84 | 1,425 | null | null | null | 15.16 | 12.37 | 16.88 | 4.94 | 1,051.53 | null | null | null | null | null | null | null | 2026-03-30T16:23:42.651140+00:00 |
sglang_core-0027 | 0 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 71.536661 | 1,091.57 | 1,016.54 | 1,825.9 | null | null | null | 15.74 | 12.23 | 15.42 | 5.14 | 1,092.78 | null | null | null | null | null | null | null | 2026-03-30T10:39:19.404150+00:00 |
sglang_core-0027 | 1 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.680576 | 864.67 | 741.56 | 1,407.91 | null | null | null | 14.62 | 12.22 | 16.21 | 5.28 | 1,123.41 | null | null | null | null | null | null | null | 2026-03-30T10:40:27.085283+00:00 |
sglang_core-0027 | 2 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.674972 | 778.59 | 773.21 | 1,324.67 | null | null | null | 14.62 | 12.23 | 15.26 | 5.3 | 1,128.03 | null | null | null | null | null | null | null | 2026-03-30T10:41:33.760798+00:00 |
sglang_core-0027 | 3 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.806051 | 778.05 | 647.93 | 1,392.33 | null | null | null | 14.94 | 12.25 | 15.45 | 5.29 | 1,124.08 | null | null | null | null | null | null | null | 2026-03-30T10:42:41.567422+00:00 |
sglang_core-0027 | 4 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 68.769496 | 786.78 | 776.99 | 1,342.67 | null | null | null | 14.69 | 12.23 | 15.52 | 5.3 | 1,126.99 | null | null | null | null | null | null | null | 2026-03-30T10:43:50.337506+00:00 |
sglang_core-0027 | 5 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.621472 | 778.42 | 648.34 | 1,391.2 | null | null | null | 14.93 | 12.25 | 15.39 | 5.29 | 1,124.42 | null | null | null | null | null | null | null | 2026-03-30T10:44:56.959705+00:00 |
sglang_core-0027 | 6 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.719464 | 774.59 | 720.8 | 1,344.27 | null | null | null | 14.73 | 12.23 | 15.25 | 5.3 | 1,127.4 | null | null | null | null | null | null | null | 2026-03-30T10:46:04.679762+00:00 |
sglang_core-0027 | 7 | sglang | claude_code | sonnet-4.5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.725703 | 774.94 | 648.53 | 1,384.29 | null | null | null | 14.93 | 12.24 | 15.93 | 5.29 | 1,124.51 | null | null | null | null | null | null | null | 2026-03-30T10:47:11.406046+00:00 |
sglang_core-0027 | 0 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.133451 | 1,011.85 | 946.49 | 1,680.27 | null | null | null | 15.13 | 12.25 | 15.57 | 5.21 | 1,108.33 | null | null | null | null | null | null | null | 2026-03-30T16:25:15.755990+00:00 |
sglang_core-0027 | 1 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.44851 | 782.54 | 715.35 | 1,325.1 | null | null | null | 14.68 | 12.24 | 15.51 | 5.28 | 1,123.48 | null | null | null | null | null | null | null | 2026-03-30T16:26:21.205045+00:00 |
sglang_core-0027 | 2 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 64.814222 | 796.84 | 736.16 | 1,340.42 | null | null | null | 14.72 | 12.21 | 15.68 | 5.27 | 1,121.49 | null | null | null | null | null | null | null | 2026-03-30T16:27:26.019801+00:00 |
sglang_core-0027 | 3 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.479872 | 788.85 | 669.36 | 1,399.8 | null | null | null | 14.99 | 12.25 | 15.24 | 5.27 | 1,119.87 | null | null | null | null | null | null | null | 2026-03-30T16:28:31.500209+00:00 |
sglang_core-0027 | 4 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.457112 | 781.14 | 681.04 | 1,398.53 | null | null | null | 15.04 | 12.26 | 15.26 | 5.26 | 1,119.37 | null | null | null | null | null | null | null | 2026-03-30T16:29:36.958132+00:00 |
sglang_core-0027 | 5 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.220241 | 771.77 | 706.01 | 1,314.47 | null | null | null | 14.67 | 12.24 | 15.5 | 5.29 | 1,124.69 | null | null | null | null | null | null | null | 2026-03-30T16:30:42.178899+00:00 |
sglang_core-0027 | 6 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.377456 | 792.2 | 651.83 | 1,377.88 | null | null | null | 14.88 | 12.24 | 15.3 | 5.27 | 1,121.09 | null | null | null | null | null | null | null | 2026-03-30T16:31:47.556892+00:00 |
sglang_core-0027 | 7 | sglang | codex_cli | gpt-5 | 6b231325b9782555eb8e1cfcf27820003a98382b | b1c8d4e9f31953560f2db45a3b6e68099ef00c13 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 64.917255 | 793.23 | 804.51 | 1,328.44 | null | null | null | 14.65 | 12.24 | 15.76 | 5.28 | 1,123.2 | null | null | null | null | null | null | null | 2026-03-30T16:32:52.474745+00:00 |
sglang_core-0033 | 0 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 68.735182 | 872.12 | 727.06 | 1,704.03 | null | null | null | 15.84 | 12.21 | 14.52 | 5.22 | 1,110.58 | null | null | null | null | null | null | null | 2026-03-30T10:52:15.015776+00:00 |
sglang_core-0033 | 1 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 320.392912 | 754.43 | 707.31 | 1,294.52 | null | null | null | 14.48 | 12.23 | 15.56 | 5.33 | 1,134.18 | null | null | null | null | null | null | null | 2026-03-30T10:57:35.409254+00:00 |
sglang_core-0033 | 2 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 197.337298 | 770.05 | 659.72 | 1,379.48 | null | null | null | 14.81 | 12.22 | 14.68 | 5.31 | 1,129.06 | null | null | null | null | null | null | null | 2026-03-30T11:00:52.747133+00:00 |
sglang_core-0033 | 3 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 197.842391 | 769.59 | 659.17 | 1,361.82 | null | null | null | 14.72 | 12.21 | 14.83 | 5.32 | 1,130.38 | null | null | null | null | null | null | null | 2026-03-30T11:04:10.590019+00:00 |
sglang_core-0033 | 4 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.969321 | 750.5 | 700.68 | 1,291.63 | null | null | null | 14.48 | 12.24 | 15.12 | 5.33 | 1,134.41 | null | null | null | null | null | null | null | 2026-03-30T11:05:17.559919+00:00 |
sglang_core-0033 | 5 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.576036 | 837.03 | 714.72 | 1,370.47 | null | null | null | 14.45 | 12.21 | 15.86 | 5.31 | 1,129.91 | null | null | null | null | null | null | null | 2026-03-30T11:06:24.136518+00:00 |
sglang_core-0033 | 6 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.081664 | 780.87 | 653.02 | 1,379 | null | null | null | 14.74 | 12.19 | 15.91 | 5.31 | 1,129.32 | null | null | null | null | null | null | null | 2026-03-30T11:07:30.218778+00:00 |
sglang_core-0033 | 7 | sglang | claude_code | sonnet-4.5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.110909 | 749.57 | 697.52 | 1,293.82 | null | null | null | 14.48 | 12.21 | 15.13 | 5.34 | 1,134.81 | null | null | null | null | null | null | null | 2026-03-30T11:08:36.330273+00:00 |
sglang_core-0033 | 0 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.669866 | 775.23 | 697.96 | 1,417.01 | null | null | null | 14.94 | 12.24 | 14.63 | 5.3 | 1,127.23 | null | null | null | null | null | null | null | 2026-03-30T16:34:34.381356+00:00 |
sglang_core-0033 | 1 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.005524 | 865.06 | 739.23 | 1,463.29 | null | null | null | 14.74 | 12.22 | 14.96 | 5.29 | 1,124.88 | null | null | null | null | null | null | null | 2026-03-30T16:35:40.387499+00:00 |
sglang_core-0033 | 2 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.636015 | 855.16 | 730.79 | 1,391.57 | null | null | null | 14.45 | 12.23 | 15.16 | 5.31 | 1,129.01 | null | null | null | null | null | null | null | 2026-03-30T16:36:46.024072+00:00 |
sglang_core-0033 | 3 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.598153 | 749.27 | 695.87 | 1,294.77 | null | null | null | 14.48 | 12.23 | 14.85 | 5.34 | 1,134.69 | null | null | null | null | null | null | null | 2026-03-30T16:37:51.622782+00:00 |
sglang_core-0033 | 4 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 70.709687 | 756.14 | 677.76 | 1,293.25 | null | null | null | 14.45 | 12.23 | 15.26 | 5.33 | 1,134.16 | null | null | null | null | null | null | null | 2026-03-30T16:39:02.333008+00:00 |
sglang_core-0033 | 5 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.234716 | 751.94 | 676.24 | 1,288.58 | null | null | null | 14.44 | 12.22 | 14.93 | 5.34 | 1,135.45 | null | null | null | null | null | null | null | 2026-03-30T16:40:07.568247+00:00 |
sglang_core-0033 | 6 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.801666 | 856.54 | 735.25 | 1,389.39 | null | null | null | 14.45 | 12.23 | 15.39 | 5.31 | 1,128.45 | null | null | null | null | null | null | null | 2026-03-30T16:41:14.370471+00:00 |
sglang_core-0033 | 7 | sglang | codex_cli | gpt-5 | 73b13e69b4207f240650c6b51eba7a7204f64939 | 8609e637a961dd0bd17bbf7f8f81b34cb2f7863a | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 64.95393 | 773.91 | 644.01 | 1,371.09 | null | null | null | 14.73 | 12.24 | 15.01 | 5.31 | 1,129.89 | null | null | null | null | null | null | null | 2026-03-30T16:42:19.324978+00:00 |
sglang_core-0047 | 0 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 100.927168 | 869.6 | 871.41 | 1,384.93 | null | null | null | 15.02 | 12.36 | 17.27 | 5.1 | 1,084.42 | null | null | null | null | null | null | null | 2026-03-30T11:12:44.154444+00:00 |
sglang_core-0047 | 1 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.648371 | 1,054.28 | 1,063.02 | 1,624.66 | null | null | null | 14.96 | 12.39 | 17.6 | 5.07 | 1,077.82 | null | null | null | null | null | null | null | 2026-03-30T11:13:49.803315+00:00 |
sglang_core-0047 | 2 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.620197 | 831.25 | 693.14 | 1,477.81 | null | null | null | 15.3 | 12.38 | 16.94 | 5.11 | 1,086.93 | null | null | null | null | null | null | null | 2026-03-30T11:14:55.424148+00:00 |
sglang_core-0047 | 3 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.965702 | 833.91 | 849.82 | 1,421.22 | null | null | null | 15.05 | 12.4 | 17.49 | 5.11 | 1,087.46 | null | null | null | null | null | null | null | 2026-03-30T11:16:02.390441+00:00 |
sglang_core-0047 | 4 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.547345 | 835.64 | 854.42 | 1,421.2 | null | null | null | 15.03 | 12.38 | 17.09 | 5.12 | 1,088.61 | null | null | null | null | null | null | null | 2026-03-30T11:17:09.938382+00:00 |
sglang_core-0047 | 5 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.457353 | 811.74 | 870.27 | 1,412.6 | null | null | null | 15.1 | 12.41 | 16.7 | 5.12 | 1,088.6 | null | null | null | null | null | null | null | 2026-03-30T11:18:16.396285+00:00 |
sglang_core-0047 | 6 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.559307 | 814.55 | 873.63 | 1,414.88 | null | null | null | 15.09 | 12.4 | 17.24 | 5.13 | 1,090.67 | null | null | null | null | null | null | null | 2026-03-30T11:19:21.956115+00:00 |
sglang_core-0047 | 7 | sglang | claude_code | sonnet-4.5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 71.316979 | 824.07 | 861.49 | 1,410.62 | null | null | null | 15.02 | 12.37 | 17.57 | 5.13 | 1,090.22 | null | null | null | null | null | null | null | 2026-03-30T11:20:33.273681+00:00 |
sglang_core-0047 | 0 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 101.449926 | 896.39 | 824.79 | 1,401.26 | null | null | null | 14.95 | 12.36 | 16.96 | 5.1 | 1,083.95 | null | null | null | null | null | null | null | 2026-03-30T16:44:32.940693+00:00 |
sglang_core-0047 | 1 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.418334 | 1,054.93 | 918.77 | 1,705.07 | null | null | null | 15.34 | 12.39 | 16.8 | 5.05 | 1,072.87 | null | null | null | null | null | null | null | 2026-03-30T16:45:40.359567+00:00 |
sglang_core-0047 | 2 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.264336 | 829.56 | 861.98 | 1,412.59 | null | null | null | 15.04 | 12.38 | 16.77 | 5.12 | 1,089.11 | null | null | null | null | null | null | null | 2026-03-30T16:46:46.624486+00:00 |
sglang_core-0047 | 3 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 65.651062 | 830.26 | 830.39 | 1,410.36 | null | null | null | 15 | 12.37 | 17 | 5.12 | 1,089.74 | null | null | null | null | null | null | null | 2026-03-30T16:47:52.276104+00:00 |
sglang_core-0047 | 4 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.909247 | 834.42 | 695.93 | 1,482.83 | null | null | null | 15.32 | 12.4 | 16.92 | 5.1 | 1,084.73 | null | null | null | null | null | null | null | 2026-03-30T16:48:59.185921+00:00 |
sglang_core-0047 | 6 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.819186 | 834.19 | 850.9 | 1,421.4 | null | null | null | 15.04 | 12.39 | 17.09 | 5.12 | 1,088.33 | null | null | null | null | null | null | null | 2026-03-30T16:51:13.898807+00:00 |
sglang_core-0047 | 7 | sglang | codex_cli | gpt-5 | a191a0e47c2f0b0c8aed28080b9cb78624365e92 | 8c7279c24e535681478188967b3007916b87b3d0 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.501129 | 923.98 | 791.44 | 1,507.03 | null | null | null | 15.02 | 12.39 | 17.78 | 5.1 | 1,083.53 | null | null | null | null | null | null | null | 2026-03-30T16:52:21.400481+00:00 |
sglang_core-0065 | 0 | sglang | claude_code | sonnet-4.5 | da47621ccc4f8e8381f3249257489d5fe32aff1b | 22a6b9fc051154347b6eb5064d2f6ef9b4dba471 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.805754 | 773.02 | 694.95 | 1,413.87 | null | null | null | 14.93 | 12.2 | 14.65 | 5.3 | 1,126.92 | null | null | null | null | null | null | null | 2026-03-30T11:25:24.040074+00:00 |
sglang_core-0065 | 1 | sglang | claude_code | sonnet-4.5 | da47621ccc4f8e8381f3249257489d5fe32aff1b | 22a6b9fc051154347b6eb5064d2f6ef9b4dba471 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.060505 | 835.6 | 711.87 | 1,371.3 | null | null | null | 14.44 | 12.2 | 15.08 | 5.31 | 1,129.49 | null | null | null | null | null | null | null | 2026-03-30T11:26:30.101185+00:00 |
sglang_core-0065 | 2 | sglang | claude_code | sonnet-4.5 | da47621ccc4f8e8381f3249257489d5fe32aff1b | 22a6b9fc051154347b6eb5064d2f6ef9b4dba471 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.500832 | 867.41 | 738.5 | 1,459.72 | null | null | null | 14.71 | 12.2 | 14.8 | 5.29 | 1,124.07 | null | null | null | null | null | null | null | 2026-03-30T11:27:37.602653+00:00 |
sglang_core-0065 | 3 | sglang | claude_code | sonnet-4.5 | da47621ccc4f8e8381f3249257489d5fe32aff1b | 22a6b9fc051154347b6eb5064d2f6ef9b4dba471 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 67.971857 | 747.13 | 671.71 | 1,280.3 | null | null | null | 14.42 | 12.19 | 15.01 | 5.34 | 1,135.39 | null | null | null | null | null | null | null | 2026-03-30T11:28:45.575107+00:00 |
sglang_core-0065 | 4 | sglang | claude_code | sonnet-4.5 | da47621ccc4f8e8381f3249257489d5fe32aff1b | 22a6b9fc051154347b6eb5064d2f6ef9b4dba471 | serving | python -m sglang.bench_serving --backend sglang --model meta-llama/Llama-3.1-8B-Instruct --num-prompts 100 | meta-llama/Llama-3.1-8B-Instruct | success | 66.155686 | 759.89 | 678 | 1,300.27 | null | null | null | 14.46 | 12.19 | 15.11 | 5.33 | 1,133.98 | null | null | null | null | null | null | null | 2026-03-30T11:29:51.731361+00:00 |
End of preview. Expand in Data Studio
ISO-Bench Pass@k GPU Benchmark Results
Agent-generated optimization patches benchmarked on real GPU hardware (NVIDIA H100 80GB).
Scope: Lossfunk/ISO-Bench — 54 tasks (39 vLLM + 15 SGLang) Patches from: Inferencebench/pass-at-k-samples
Summary
| vLLM | SGLang | Total | |
|---|---|---|---|
| Tasks benchmarked | 29/39 | 10/15 | 39/54 |
| Successful benchmarks | 444 | 150 | 594 |
| Agent | vLLM | SGLang | Total |
|---|---|---|---|
| Claude Code (Sonnet 4.5) | 230 | 76 | 306 |
| Codex CLI (GPT-5) | 214 | 74 | 288 |
Schema
| Column | Type | Description |
|---|---|---|
item_id |
str | ISO-Bench task ID (e.g., vllm_core-0000) |
sample_index |
int | Pass@k sample index (0-7) |
repo |
str | vllm or sglang |
agent_name |
str | claude_code or codex_cli |
agent_model |
str | sonnet-4.5 or gpt-5 |
human_commit |
str | Human optimization commit hash |
parent_commit |
str | Baseline (pre-optimization) commit hash |
benchmark_mode |
str | serving, standalone, prefix_caching, offline |
llm_model |
str | LLM model benchmarked |
duration_s |
float | Total benchmark duration (seconds) |
Serving Metrics
ttft_mean_ms, ttft_median_ms, ttft_p99_ms — Time To First Token
tpot_mean_ms, tpot_median_ms, tpot_p99_ms — Time Per Output Token
itl_mean_ms, itl_median_ms, itl_p99_ms — Inter-Token Latency
request_throughput_req_s, output_token_throughput_tok_s, total_token_throughput_tok_s
Latency Metrics
latency_avg_ms, latency_p50_ms, latency_p99_ms
Throughput Metrics
throughput_tok_s, elapsed_time_s, input_throughput_tok_s
Usage
from datasets import load_dataset
ds = load_dataset("Inferencebench/iso-bench-pass-at-k-results", split="train")
# Filter by repo and agent
vllm_cc = ds.filter(lambda x: x["repo"] == "vllm" and x["agent_name"] == "claude_code")
sglang_cx = ds.filter(lambda x: x["repo"] == "sglang" and x["agent_name"] == "codex_cli")
# Get results for a specific task
task = ds.filter(lambda x: x["item_id"] == "vllm_core-0005")
Hardware & Method
- GPU: NVIDIA H100 80GB PCIe
- vLLM: Docker containers (
shikhar481/vllm_fixed_human_images) - SGLang: Docker containers (
ayushnangia16/nvidia-sglang-docker) - Execution: Persistent container per task, sequential pass@k samples, shared setup across agents
- Date: 2026-03-29 to 2026-03-31
- Downloads last month
- 16