List your company

Multi-turn Dialogue on MT-Bench (Token/s and Speedup)

270.96Tokens/sEVICT
92.1216138.5508184.98231.4092May 1, 2026

Evaluation results

MethodLinks
EVICTTemperature=0, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=4, steps=4, topk=8, draft_tokens=322026.05270.96—
EVICTTemperature=1, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=4, steps=4, topk=8, draft_tokens=322026.05254.83—
EAGLE-3Temperature=0, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=4, steps=4, topk=8, draft_tokens=322026.05221.97—
EAGLE-3Temperature=1, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=4, steps=4, topk=8, draft_tokens=322026.05213.12—
DDDTemperature=0, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=4, steps=4, topk=8, draft_tokens=322026.05212.33—
DDDTemperature=1, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=4, steps=4, topk=8, draft_tokens=322026.05203.17—
EAGLE-3 + EVICTTemperature=0, Backbone=Qwen3-30B-A3B2026.05190.34—
EAGLE-3 + EVICTTemperature=1, Backbone=Qwen3-30B-A3B2026.05180.18—
VanillaTemperature=0, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=42026.05174.57—
EAGLE-3Temperature=0, Backbone=Qwen3-30B-A3B2026.05171.08—
VanillaTemperature=1, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=42026.05170.29—
EAGLE-3Temperature=1, Backbone=Qwen3-30B-A3B2026.05162.26—
EAGLE-3 + DDDTemperature=0, Backbone=Qwen3-30B-A3B2026.05159.16—
EAGLE-3 + DDDTemperature=1, Backbone=Qwen3-30B-A3B2026.05152.31—
VanillaTemperature=0, Backbone=Qwen3-30B-A3B2026.05143.1—
VanillaTemperature=1, Backbone=Qwen3-30B-A3B2026.05140.13—
LookaheadTemperature=0, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=42026.05133.59—
LookaheadTemperature=1, Backbone=Ling-flash-2.0, Batch Size=1, TP (Tensor Parallelism)=42026.05129.01—
LookaheadTemperature=0, Backbone=Qwen3-30B-A3B2026.05101.59—
LookaheadTemperature=1, Backbone=Qwen3-30B-A3B2026.0599—