MiMo-V2.6-Flash-RL
70 records across 9 hardware.
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 4x AMD MI210 | MiMo-V2.6-Flash-RL 309B | mxfp4 | vLLM | 160 | – | – | – | – | 2026-09-29 | community | – | source |
| M5 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 104.8 | 3207.0 | – | – | 4096 | 2026-10-08 | community | – | source |
| M5 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 92.9 | 3381.0 | – | – | 8192 | 2026-10-08 | community | – | source |
| M5 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 84.9 | 1844.0 | – | – | 1024 | 2026-10-08 | community | – | source |
| M5 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 71.4 | 552.0 | – | – | 1024 | 2026-10-07 | community | – | source |
| M5 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 68.6 | 2329.0 | – | – | 32768 | 2026-10-08 | community | – | source |
| 4x CMP 170HX | MiMo-V2.6-Flash-RL | FP8 | vLLM | 63 | – | – | – | 5120 | 2026-09-28 | community | – | source |
| 4x CMP 170HX | MiMo-V2.6-Flash-RL | FP8 | vLLM | 56.1 | – | – | – | 67000 | 2026-09-28 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 55.5 | 427.9 | – | – | 1024 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 54.9 | 636.6 | – | – | 4096 | 2026-10-12 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 54.5 | 484.4 | – | – | 4096 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 54.3 | 621.0 | – | – | 8192 | 2026-10-12 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 53.6 | 447.2 | – | – | 16384 | 2026-10-16 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 53.3 | 617.7 | – | – | 8192 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 53.3 | 482.2 | – | – | 4096 | 2026-10-16 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 53.1 | 424.0 | – | – | 1024 | 2026-10-16 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 52.8 | 473.1 | – | – | 8192 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 51.7 | 409.1 | – | – | 1024 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 51.7 | 673.1 | – | – | 4096 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 51.5 | 480.7 | – | – | 4096 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 51.2 | 531.9 | – | – | 1024 | 2026-10-12 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 51.0 | 585.5 | – | – | 16384 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 50.9 | 606.1 | – | – | 1024 | 2026-10-10 | community | – | source |
| M2 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 50.2 | 344.0 | – | – | 1024 | 2026-10-17 | community | – | source |
| M2 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 50.2 | 344.0 | – | – | 1024 | 2026-10-17 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 50.1 | 450.1 | – | – | 16384 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 50.1 | 632.5 | – | – | 4096 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 49.4 | 644.3 | – | – | 8192 | 2026-10-16 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 48.7 | 655.5 | – | – | 8192 | 2026-10-10 | community | contradiction | source |
| M2 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 48.5 | 375.5 | – | – | 4096 | 2026-10-17 | community | – | source |
| M2 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 48.5 | 375.5 | – | – | 4096 | 2026-10-17 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 47.7 | 471.5 | – | – | 8192 | 2026-10-07 | community | – | source |
| M5 Max | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 47.4 | 1293.0 | – | – | 1024 | 2026-10-14 | community | – | source |
| M5 Max | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 47.4 | 579.6 | – | – | 1024 | 2026-10-13 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 47.1 | 587.4 | – | – | 16384 | 2026-10-12 | community | – | source |
| M5 Max | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 46.4 | 1300.0 | – | – | 4096 | 2026-10-13 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 46.2 | 522.0 | – | – | 32768 | 2026-10-10 | community | – | source |
| M5 Max | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 46.0 | 1753.0 | – | – | 8192 | 2026-10-13 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 46.0 | 607.0 | – | – | 16384 | 2026-10-16 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 45.6 | 618.3 | – | – | 16384 | 2026-10-10 | community | – | source |
| M5 Max | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 44.9 | 1461.0 | – | – | 16384 | 2026-10-13 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | oQ8e | oMLX | 44.7 | 573.2 | – | – | 1024 | 2026-10-14 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | oQ8e | oMLX | 44.7 | 573.2 | – | – | 1024 | 2026-10-14 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 44.2 | 402.8 | – | – | 32768 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | oQ8e | oMLX | 44.1 | 644.6 | – | – | 4096 | 2026-10-14 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | oQ8e | oMLX | 44.1 | 644.6 | – | – | 4096 | 2026-10-14 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 43.9 | 646.2 | – | – | 4096 | 2026-10-16 | community | contradiction | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 43.5 | 545.0 | – | – | 1024 | 2026-10-10 | community | – | source |
| 4x CMP 170HX | MiMo-V2.6-Flash-RL | FP8 | vLLM | 42.5 | – | – | – | 5120 | 2026-09-28 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 41.4 | 547.6 | – | – | 32768 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 41.3 | 447.7 | – | – | 16384 | 2026-10-07 | community | contradiction | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 40.9 | 389.3 | – | – | 32768 | 2026-10-07 | community | – | source |
| M2 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 40.7 | 325.8 | – | – | 32768 | 2026-10-18 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 37.1 | 330.3 | – | – | 65536 | 2026-10-07 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 34.4 | 439.8 | – | – | 65536 | 2026-10-10 | community | – | source |
| 4x CMP 170HX | MiMo-V2.6-Flash-RL | FP8 | vLLM | 33.7 | – | – | – | 67000 | 2026-09-28 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 33.0 | 304.0 | – | – | 65536 | 2026-10-07 | community | contradiction | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 31.4 | 424.2 | – | – | 65536 | 2026-10-10 | community | – | source |
| 4x AMD MI210 | MiMo-V2.6-Flash-RL 309B | mxfp4 | vLLM | 31 | – | – | – | – | 2026-09-29 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 28.6 | 243.7 | – | – | 131072 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 25.4 | 309.8 | – | – | 131072 | 2026-10-10 | community | – | source |
| Radeon 8065S | MiMo-V2.6-Flash-RL llama-bench -fa 1 -b 2048 -ub 2048, 3 runs; ROCmFPX main (fused gate/up expert tensors, from PR #33), Mesa RADV | MXFP4-Q5K mixed (4.29 bpw) | llama.cpp (Vulkan) | 23.6 | – | – | – | tg128 | 2026-10-02 | community | – | source |
| Radeon 8065S | MiMo-V2.6-Flash-RL llama-bench -fa 1 -b 2048 -ub 2048, 3 runs; ROCmFPX main (fused gate/up expert tensors, from PR #33), Mesa RADV | MXFP4-Q5K mixed (4.29 bpw) | llama.cpp (Vulkan) | 22.1 | – | – | – | 32768 | 2026-10-02 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | 4bit | oMLX | 20.5 | 233.1 | – | – | 199680 | 2026-10-10 | community | – | source |
| M3 Ultra | MiMo-V2.6-Flash-RL | mxfp4 | oMLX | 19.1 | 220.3 | – | – | 199680 | 2026-10-10 | community | – | source |
| 8× NVIDIA RTX 3080 20GB | MiMo-V2.6-Flash-RL | MXFP4 | llama.cpp | 18.6 | – | – | – | 262144 | 2026-09 | community | – | source |
| 4x CMP 170HX | MiMo-V2.6-Flash-RL | FP8 | vLLM | 16.5 | – | – | – | 67000 | 2026-09-28 | community | – | source |
| Radeon 8060S | MiMo-V2.6-Flash-RL 309B | IQ2 | llama.cpp (Vulkan) | 2.6 | – | – | – | 32768 | 2026-10-05 | community | – | source |
| Radeon 8065S | MiMo-V2.6-Flash-RL llama-bench -fa 1 -b 2048 -ub 2048, 3 runs; ROCmFPX main (fused gate/up expert tensors, from PR #33), Mesa RADV | MXFP4-Q5K mixed (4.29 bpw) | llama.cpp (Vulkan) | – | 484.9 | – | – | – | 2026-10-02 | community | – | source |
| Radeon 8065S | MiMo-V2.6-Flash-RL llama-bench -fa 1 -b 2048 -ub 2048, 3 runs; ROCmFPX main (fused gate/up expert tensors, from PR #33), Mesa RADV | MXFP4-Q5K mixed (4.29 bpw) | llama.cpp (Vulkan) | – | 258.7 | – | – | 32768 | 2026-10-02 | community | – | source |