Gemma 4 E2B

model page · concurrency sweep 1–24 agents

Gemma 4 E2B

server model id: google/gemma-4-e2b
thinkingKV Q8_0ctx 32k2B params

Near the bottom of the field — 72 tok/s solo, 293 tok/s peak, ~99% of output hidden reasoning under load, and the second-worst TTFT after Qwen3 4B (11.8s solo, up to 59.8s).

Key findings

  • One of the slowest profiles tested: only 72 tok/s solo (58 visible), and combined all-token throughput peaks at just 293 tok/s at 24 agents (4.1x) — near the bottom of the field alongside qwen3-4b.
  • Near-total reasoning collapse under load: visible content drops to 2–6 tok/s at 20–24 agents (~99% of output is hidden reasoning_content). At 21–24 agents agents produce essentially zero usable content in the 60s window.
  • Extreme TTFT — the second-worst after qwen3-4b: 11.8s solo, rising steadily to 41–58s under load, with max values up to 59.8s. Agents routinely wait almost the whole run for their first token.
  • Steady but slow scaling: all-token throughput climbs monotonically 72 → 293 without rolloff, but absolute numbers are low.
  • Errors: 0 across all 24 runs.

Sweep — total concurrency 1–24

agentswall (s)contentreasontotalcomb tok/scomb allper-agentper-agent allscale %TTFT meanTTFT maxoktimeouterr
160.03,4878144,301587272.372.3100%11.8s11.8s010
260.05,3831,7617,1449011958.659.8165%14.1s14.2s020
360.05,9822,5288,51010014246.348.3197%16.9s18.2s030
460.06,5713,77610,34710917242.443.7239%21.3s27.8s040
560.06,1184,25610,37410217334.335.2240%24.3s25.9s050
660.06,7995,06211,86111319832.533.8275%25.1s26.0s060
760.06,1525,89412,04610220128.129.3279%28.7s31.5s070
860.06,1796,74912,92810321526.427.6299%30.8s40.7s080
960.05,7027,77913,4819522524.425.7312%34.1s42.8s090
1060.05,9458,51314,4589924123.524.9335%34.7s41.7s0100
1160.05,2469,19514,4418724121.622.6335%37.9s42.9s0110
1260.04,74110,38915,1307925220.821.8350%41.0s48.7s0120
1360.04,41110,75315,1647325319.120.2351%42.3s54.3s0130
1460.04,73211,44916,1817927019.120.1375%42.3s53.0s0140
1560.03,69312,50516,1986227016.718.7375%45.3s55.9s0150
1660.03,38713,52116,9085628217.418.4392%47.8s56.5s0160
1760.01,86914,53816,4073127314.216.8379%52.2s58.9s0170
1860.02,31114,85817,1693828614.016.7397%50.8s58.4s0180
1960.01,54415,59217,1362628510.515.9396%52.2s57.1s0190
2060.085216,54617,398142908.915.3403%55.1s59.2s0200
2160.015716,35216,50932753.713.9382%57.8s59.7s0210
2260.034816,72817,07662843.213.7394%54.7s59.8s0220
2360.09517,25517,35022890.613.4401%52.5s52.5s0230
24 peak60.023517,37217,60742930.513.1407%41.4s41.4s0240

* burst artifact — agents queued ~11–16s then generated in a short burst; the per-agent timer inflates the number. Trust combined throughput and TTFT columns. Peak row = highest combined (all-token) throughput.

Charts

Whole sweep at a glance — combined throughput, scaling efficiency, per-agent throughput and outcomes across every concurrency level.
Whole sweep at a glance — combined throughput, scaling efficiency, per-agent throughput and outcomes across every concurrency level.
Combined throughput (all tokens) vs concurrent agents. Annotated peak with the concurrency level where it occurred.
Combined throughput (all tokens) vs concurrent agents. Annotated peak with the concurrency level where it occurred.
Per-agent throughput erosion as slots contend. Red stars mark burst-timing artifacts (queued ≥ 10s then generated in a burst — inflated values, not real throughput).
Per-agent throughput erosion as slots contend. Red stars mark burst-timing artifacts (queued ≥ 10s then generated in a burst — inflated values, not real throughput).
Scaling efficiency relative to the single-agent baseline (100%).
Scaling efficiency relative to the single-agent baseline (100%).
Time to first token — mean (solid) and max (dashed). Log scale. Gaps mean no visible content ever arrived within the 60s window.
Time to first token — mean (solid) and max (dashed). Log scale. Gaps mean no visible content ever arrived within the 60s window.
Total tokens per 60s run — all tokens vs visible content only.
Total tokens per 60s run — all tokens vs visible content only.
How each run of N agents finished: completed, timed out (60s budget), or errored.
How each run of N agents finished: completed, timed out (60s budget), or errored.
The concurrency trade-off on one chart — combined throughput rises while per-agent speed falls.
The concurrency trade-off on one chart — combined throughput rises while per-agent speed falls.
Visible content vs hidden reasoning tokens per run, plus the reasoning share (%).
Visible content vs hidden reasoning tokens per run, plus the reasoning share (%).

Raw data & downloads