Gemma 4 E4B
- Input
- $0.02/M tok
- Output
- $0.10/M tok
- Blended
- $0.04/M tok · 3:1
- Speed
- 30tok/s
- Context
- 128Ktokens
Long-tail import from Artificial Analysis (2026-09-24); AA variant: Gemma 4 E4B (Non-reasoning). Price = AA median host price.
Rankings
Writing & chat
Verified#113 / 162rank by quality26.9±9.5quality / 1001/5 benchmarks measuredPrice frontierVision
Verified#83 / 98rank by quality24.7±8.2quality / 1001/6 benchmarks measuredPrice frontierHard reasoning
Verified#167 / 218rank by quality21.1±3.7quality / 1003/4 benchmarks measuredPrice frontierAgents
Verified#134 / 187rank by quality18.2±10.0quality / 1001/7 benchmarks measuredPrice frontier
Self-hostingEstimated
How we estimate →What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.
- Hardware
- 1× B200 180GB
- BF16 weights
- Throughput
- ~13,802 tok/s
- many requests batched
- Price
- $0.054–0.48 /M tok
- busy → light use
- On your own machine
- RTX 5090 32GB (int4, 300+ tok/s single-stream)
- single consumer GPU or Mac
Assumptions (5)
- BF16 weights (8B params, 4.5B active per token) + 25% KV-cache headroom ≈ 20 GB
- 1× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
- ~13,802 output tok/s aggregate at batch 256 (bandwidth-bound); MoE compute scales with active params
- Blended 3:1 input:output; prefill ~118,125 tok/s
- Low = 75% utilization at the low GPU price; high = 20% at the high price
Benchmark scores
~ italic, dashed = no published score yet, estimated from related benchmarks.
Writing & chat
IFBenchVerified
#88 of 13940.5%
- #1MiniMax-M382.9%
- #2Qwen3.8 2.4T A95B82.8%
- #87Olmo 3 7B Think41.5%
- #88Gemma 4 E4B40.5%
- #89Qwen3 Next 80B A3B Instruct39.7%
artificialanalysis.ai · 2026-09-24
MMMLUEstimated
~last of 5 (estimated)below measured range (<81.3%)
No published score — estimated from related benchmarks
LMArena Text, Non-English (Elo)Estimated
~#47 of 57 (estimated)~1328 Elo
- #1Gemini 4 ArgonFew votes1512
- #2Claude Opus 5.51497
- #46gpt-oss-120b1338
- ~#47Gemma 4 E4BEstimated1328
- #47INTELLECT-3Few votes1327
No published score — estimated from related benchmarks
LMArena Text (Elo)Estimated
~#48 of 57 (estimated)~1350 Elo
- #1Gemini 4 Argon1525
- #2Claude Opus 5.51507
- #47gpt-oss-120b1352
- ~#48Gemma 4 E4BEstimated1350
- #48Llama 3.1 Nemotron Ultra 253B v1Few votes1348
No published score — estimated from related benchmarks
EQ-Bench Creative Writing v3 (Elo)Estimated
~#23 of 26 (estimated)~1402 Elo
- #1GPT-6 Astra2173
- #2Claude Fable 5.12162
- #22Mistral Large 31412
- ~#23Gemma 4 E4BEstimated1402
- #23Gemma 4 31B1368
No published score — estimated from related benchmarks
Vision
MMMU-ProVerified
#72 of 8751.2%
- #1Claude Opus 5.587.7%
- #2GPT-6 Astra86.9%
- #70NVIDIA Nemotron Nano 12B v2 VL52.9%
- #72Gemma 4 E4B51.2%
- #73Ministral 3 14B49.8%
artificialanalysis.ai · 2026-09-24
CharXiv ReasoningEstimated
~last of 12 (estimated)below measured range (<74.5%)
No published score — estimated from related benchmarks
OmniDocBench v1.5Estimated
~last of 8 (estimated)below measured range (<75.8%)
No published score — estimated from related benchmarks
GDP.pdf (AA)Estimated
~#46 of 48 (estimated)~0.2%
- #1Muse Wonder Full32.6%
- #2GPT-6 Astra31.0%
- #45Ministral 3 8B0.8%
- ~#46Gemma 4 E4BEstimated0.2%
- #46Ministral 3 3B0.0%
No published score — estimated from related benchmarks
ChartographyEstimated
~last of 25 (estimated)below measured range (<9.0%)
No published score — estimated from related benchmarks
Hard reasoning
GPQA DiamondVerified
#147 of 19254.9%
- #1GPT-6 Astra96.1%
- #2Gemini 3.8 Flash95.3%
- #146NVIDIA Nemotron Nano 9B V255.7%
- #147Gemma 4 E4B54.9%
- #148Olmo 3.1 32B Instruct53.9%
artificialanalysis.ai · 2026-09-24
Humanity's Last ExamVerified
#171 of 1984.8%
- #1Claude Fable 5.274.3%
- #2Claude Opus 5.561.4%
- #169NVIDIA Nemotron 3 Nano 4B4.9%
- #171Gemma 4 E4B4.8%
- #171Nemotron 3 Nano Omni 30B A3B Reasoning4.8%
artificialanalysis.ai · 2026-09-24
AIME (latest)Estimated
~#23 of 34 (estimated)~43.9%
- #1GLM-5.3 NVFP499.5%
- #2Beam 1.0 501B A23B97.7%
- #22Phi-4-reasoning-plus78.0%
- ~#23Gemma 4 E4BEstimated43.9%
- #23Mistral Large 338.0%
No published score — estimated from related benchmarks
Agents
τ²-bench (Telecom)Verified
#69 of 11926.0%
- #1Step 3.7 Flash98.5%
- #2DeepSeek-V4-Pro-0813-nvfp4-DSpark98.3%
- #66Ministral 3 8B26.6%
- #69Gemma 4 E4B26.0%
- #70K2 Think V225.4%
artificialanalysis.ai · 2026-09-24
OSWorld-VerifiedEstimated
~#12 of 15 (estimated)~53.0%
- #1Kimi K384.8%
- #2Qwen3.8 27B84.3%
- #11Muse Glimmer65.9%
- ~#12Gemma 4 E4BEstimated53.0%
- #12Nemotron 3.5 Super VL NVFP452.3%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~#3 of 5 (estimated)~53.8%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#21 of 22 (estimated)~40.4%
- #1Atria Dawn Preview92.5%
- #2GPT-6 Astra91.5%
- #20Nemotron 3 Ultra44.4%
- ~#21Gemma 4 E4BEstimated40.4%
- #21Mistral Small 421.3%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Estimated
~#61 of 92 (estimated)~679 Elo
- #1Claude Opus 5.51866
- #2Muse Wonder Full1857
- #60MiniCPM5-2B694
- ~#61Gemma 4 E4BEstimated679
- #61Nemotron 3.5 Lightning624
No published score — estimated from related benchmarks
AutomationBench-AAEstimated
~#79 of 85 (estimated)~0.2%
- #1Claude Fable 5.276.4%
- #2Muse Wonder Full71.3%
- #77Granite 4.2 3B0.4%
- ~#79Gemma 4 E4BEstimated0.2%
- #79gpt-oss-120b0.2%
No published score — estimated from related benchmarks
τ-Bench Banking (AA)Estimated
~#86 of 91 (estimated)~4.4%
- #1Muse Spark 1.350.5%
- #2GLM-5.350.3%
- #84Magistral Small 1.24.5%
- ~#86Gemma 4 E4BEstimated4.4%
- #86Qwen3.5 4B4.3%
No published score — estimated from related benchmarks