Gemma 4 E4B Benchmarks, Specifications & Availability

Gemma 4 E4B

Explore Gemma 4 E4B from Google: published specifications, source-linked vendor benchmarks, independent evaluator coverage and recorded pricing when available.

Published specifications

Provider: Google
Access: Open weights
License: Apache-2.0
Context window: 128K
Total parameters: 8B
Active parameters: Not published
Released: 2026-03-31
Modalities: text, image, video, audio
Family: Gemma 4

Hugging Face · Announcement · Website

Model notes

Public ungated Apache-2.0 instruction-tuned checkpoint; PT, assistant/MTP and QAT builds are not extra curated models. Vendor architecture size is retained separately from HF tensor totals including modality modules. Effective dense size is 4.5B; the larger total includes per-layer lookup embeddings. Effective is not a MoE active count. Audio input is supported for up to30 seconds; video uses frames (up to60 seconds at1fps), text output only. Context is the published configured limit, not a universal guarantee of whole-window accuracy. Reasoning is an on/off mode, not an API effort level. Report Table9 disables thinking; vision uses1120 or280 visual-token budgets. OmniDocBench uses native average edit distance, not accuracy. Card/report audio aggregates and the older launch τ² figures differ and remain source-specific observations. Official release notes date31March; evaluator release dates may say2April and do not override the vendor date. No exact public OpenRouter route found19October.

Compare Gemma 4 E4B with other models →Compare subscription plans Explore data coverage

Pricing · OpenRouter

No OpenRouter pricing is recorded for this model. Missing rates are not free.

Official / vendor benchmarks

Default headline records. Own-vendor, peer-vendor and third-party provenance remain visible in evidence; configurations may differ.

Official / vendor headline scores; expand evidence for every record
Benchmark / evaluator	Headline score	Evidence
MMLU-Pro	69.4%unknown effort	All 1 recorded result & sources 69.4% · raw 69.4 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
AIME 2026	42.5%unknown effort	All 1 recorded result & sources 42.5% · raw 42.5 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. No tools. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
LiveCodeBench v6	52%unknown effort	All 1 recorded result & sources 52% · raw 52 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
Codeforces Rating	940unknown effort	All 1 recorded result & sources 940 · raw 940 rating Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
SciCode (Gemma report; aggregation unspecified)	24%unknown effort	All 1 recorded result & sources 24% · raw 24 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
GPQA Diamond	58.6%unknown effort	All 1 recorded result & sources 58.6% · raw 58.6 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
BigBench Extra Hard (micro average)	33.1%unknown effort	All 1 recorded result & sources 33.1% · raw 33.1 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
IFBench (Gemma report protocol)	44%unknown effort	All 1 recorded result & sources 44% · raw 44 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. IFBench strict/loose convention is not stated in this table. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
IFEval	96.7%unknown effort	All 1 recorded result & sources 96.7% · raw 96.7 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
MMMLU	76.6%unknown effort	All 1 recorded result & sources 76.6% · raw 76.6 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
MRCR v2 (8 needles,128K average)	25.4%unknown effort	All 1 recorded result & sources 25.4% · raw 25.4 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
Terminal Bench Hard (Gemma report)	8%unknown effort	All 1 recorded result & sources 8% · raw 8 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
τ²-bench (Airline)	52%unknown effort	All 1 recorded result & sources 52% · raw 52 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
τ²-bench (Retail)	57.5%unknown effort	All 2 recorded results & sources 67.1% · raw 67.1 % Alternative · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. Demoted19October: own launch figure has a different retail result; retain this report observation without averaging. https://arxiv.org/pdf/2607.02770#page=5 57.5% · raw 57.5 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google launch benchmark figure; τ² Retail printed57.5. Own-launch tiebreak over later report values; source disagreement is retained explicitly. https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/
τ²-bench (Telecom)	18.4%unknown effort	All 1 recorded result & sources 18.4% · raw 18.4 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google final instruction-tuned report Table5; thinking enabled, no effort tier inferred. Report dated19June2026, not execution date. Visually checked numeric table. https://arxiv.org/pdf/2607.02770#page=5
MMMU Pro (no tools)	52.6%unknown effort	All 1 recorded result & sources 52.6% · raw 52.6 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table6; thinking,1120 visual tokens; execution date and exact output budget unpublished. Native %. https://arxiv.org/pdf/2607.02770#page=6
MMMU Pro (Gemma,280 visual tokens)	51.4%unknown effort	All 1 recorded result & sources 51.4% · raw 51.4 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table12; thinking,280 visual tokens; separate visual-resolution configuration, not a new checkpoint. https://arxiv.org/pdf/2607.02770#page=17
MATH-Vision	59.5%unknown effort	All 1 recorded result & sources 59.5% · raw 59.5 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table6; thinking,1120 visual tokens; execution date and exact output budget unpublished. Native %. https://arxiv.org/pdf/2607.02770#page=6
MATH-Vision (Gemma,280 visual tokens)	59.2%unknown effort	All 1 recorded result & sources 59.2% · raw 59.2 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table12; thinking,280 visual tokens; separate visual-resolution configuration, not a new checkpoint. https://arxiv.org/pdf/2607.02770#page=17
MedXpertQA MM	28.7%unknown effort	All 1 recorded result & sources 28.7% · raw 28.7 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table6; thinking,1120 visual tokens; execution date and exact output budget unpublished. Native %. https://arxiv.org/pdf/2607.02770#page=6
MedXpertQA MM (Gemma,280 visual tokens)	28.7%unknown effort	All 1 recorded result & sources 28.7% · raw 28.7 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table12; thinking,280 visual tokens; separate visual-resolution configuration, not a new checkpoint. https://arxiv.org/pdf/2607.02770#page=17
InfographicVQA	70%unknown effort	All 1 recorded result & sources 70% · raw 70 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table6; thinking,1120 visual tokens; execution date and exact output budget unpublished. Native %. https://arxiv.org/pdf/2607.02770#page=6
InfographicVQA (Gemma,280 visual tokens)	54.8%unknown effort	All 1 recorded result & sources 54.8% · raw 54.8 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table12; thinking,280 visual tokens; separate visual-resolution configuration, not a new checkpoint. https://arxiv.org/pdf/2607.02770#page=17
OmniDocBench1.5 (average edit distance)	0.181 edit distanceunknown effort	All 1 recorded result & sources 0.181 edit distance · raw 0.181 edit distance Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table6; thinking,1120 visual tokens; execution date and exact output budget unpublished. Native edit distance. https://arxiv.org/pdf/2607.02770#page=6
OmniDocBench1.5 (average edit distance) (Gemma,280 visual tokens)	0.307 edit distanceunknown effort	All 1 recorded result & sources 0.307 edit distance · raw 0.307 edit distance Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table12; thinking,280 visual tokens; separate visual-resolution configuration, not a new checkpoint. https://arxiv.org/pdf/2607.02770#page=17
RULER32K (Gemma,no thinking)	95.2%none effort	All 1 recorded result & sources 95.2% · raw 95.2 % Headline · none effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table9, thinking disabled. RULER32K. Native %. https://arxiv.org/pdf/2607.02770#page=8
RULER128K (Gemma,no thinking)	86.6%none effort	All 1 recorded result & sources 86.6% · raw 86.6 % Headline · none effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table9, thinking disabled. RULER128K. Native %. https://arxiv.org/pdf/2607.02770#page=8
LOFT text retrieval128K (Recall@k) (Gemma,no thinking)	58.5%none effort	All 1 recorded result & sources 58.5% · raw 58.5 % Headline · none effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table9, thinking disabled. LOFT text retrieval128K (Recall@k). Native %. https://arxiv.org/pdf/2607.02770#page=8
GraphWalks <128K (F1) (Gemma,no thinking)	50.9%none effort	All 1 recorded result & sources 50.9% · raw 50.9 % Headline · none effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table9, thinking disabled. GraphWalks <128K (F1). Native %. https://arxiv.org/pdf/2607.02770#page=8
MTOB English→Kalamang ~128K half-book (Gemma,no thinking)	37.8 chrFnone effort	All 1 recorded result & sources 37.8 chrF · raw 37.8 chrF Headline · none effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table9, thinking disabled. MTOB English→Kalamang ~128K half-book. Native chrF. https://arxiv.org/pdf/2607.02770#page=8
MTOB Kalamang→English ~128K half-book (Gemma,no thinking)	34.6 chrFnone effort	All 1 recorded result & sources 34.6 chrF · raw 34.6 chrF Headline · none effort · Own vendor Source/record date: 2026-10-19 Harness: Google report Table9, thinking disabled. MTOB Kalamang→English ~128K half-book. Native chrF. https://arxiv.org/pdf/2607.02770#page=8
τ²-bench (Tau-2)	42.2%unknown effort	All 1 recorded result & sources 42.2% · raw 42.2 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google official card; published average over3 domains, thinking on. Not recomputed from report domains. https://huggingface.co/google/gemma-4-E4B-it/blob/ee0ef6023621cff504d758262d4e04895a5af4a2/README.md
CoVoST (Gemma card aggregate)	35.54 BLEUunknown effort	All 1 recorded result & sources 35.54 BLEU · raw 35.54 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google card audio aggregate. CoVoST translation corpus BLEU.12B excludes Chinese. Distinct from report Table7/8. https://huggingface.co/google/gemma-4-E4B-it/blob/ee0ef6023621cff504d758262d4e04895a5af4a2/README.md
FLEURS (Gemma card aggregate error)	8.00%unknown effort	All 1 recorded result & sources 8.00% · raw 8 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google card audio aggregate. FLEURS error fraction converted to percent; WER/CER language conventions are not restated on card.12B excludes Chinese. Distinct from report Table7/8. https://huggingface.co/google/gemma-4-E4B-it/blob/ee0ef6023621cff504d758262d4e04895a5af4a2/README.md
CoVoST ja→English (Gemma report)	25.5 BLEUunknown effort	All 1 recorded result & sources 25.5 BLEU · raw 25.5 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; ja source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST de→English (Gemma report)	42 BLEUunknown effort	All 1 recorded result & sources 42 BLEU · raw 42 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; de source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST fr→English (Gemma report)	41 BLEUunknown effort	All 1 recorded result & sources 41 BLEU · raw 41 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; fr source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST es→English (Gemma report)	44.8 BLEUunknown effort	All 1 recorded result & sources 44.8 BLEU · raw 44.8 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; es source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST it→English (Gemma report)	43 BLEUunknown effort	All 1 recorded result & sources 43 BLEU · raw 43 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; it source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST ru→English (Gemma report)	49.4 BLEUunknown effort	All 1 recorded result & sources 49.4 BLEU · raw 49.4 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; ru source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST zh→English (Gemma report)	21.9 BLEUunknown effort	All 1 recorded result & sources 21.9 BLEU · raw 21.9 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; zh source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
CoVoST 7-language average (Gemma report)	38.2 BLEUunknown effort	All 1 recorded result & sources 38.2 BLEU · raw 38.2 BLEU Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; average source language; native corpus BLEU, not percent accuracy. https://arxiv.org/pdf/2607.02770#page=7
FLEURS en (WER,Gemma report)	6.50%unknown effort	All 1 recorded result & sources 6.50% · raw 6.5 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS ko (CER,Gemma report)	5.30%unknown effort	All 1 recorded result & sources 5.30% · raw 5.3 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription CER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS ja (CER,Gemma report)	7.80%unknown effort	All 1 recorded result & sources 7.80% · raw 7.8 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription CER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS de (WER,Gemma report)	6.10%unknown effort	All 1 recorded result & sources 6.10% · raw 6.1 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS fr (WER,Gemma report)	8.00%unknown effort	All 1 recorded result & sources 8.00% · raw 8 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS hi (WER,Gemma report)	8.60%unknown effort	All 1 recorded result & sources 8.60% · raw 8.6 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS es (WER,Gemma report)	3.50%unknown effort	All 1 recorded result & sources 3.50% · raw 3.5 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS it (WER,Gemma report)	3.20%unknown effort	All 1 recorded result & sources 3.20% · raw 3.2 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS pt-br (WER,Gemma report)	4.60%unknown effort	All 1 recorded result & sources 4.60% · raw 4.6 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS ru (WER,Gemma report)	6.80%unknown effort	All 1 recorded result & sources 6.80% · raw 6.8 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS ar (WER,Gemma report)	16.20%unknown effort	All 1 recorded result & sources 16.20% · raw 16.2 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription WER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS zh (CER,Gemma report)	13.60%unknown effort	All 1 recorded result & sources 13.60% · raw 13.6 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription CER. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
FLEURS average (mixed WER/CER macro error,Gemma report)	7.50%unknown effort	All 1 recorded result & sources 7.50% · raw 7.5 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: Google report audio Table7; transcription mixed WER/CER macro error. Source fraction retained. https://arxiv.org/pdf/2607.02770#page=7
MMMLU (Gemma launch figure label)	69.4%unknown effort	All 1 recorded result & sources 69.4% · raw 69.4 % Headline · unknown effort · Own vendor Source/record date: 2026-10-19 Harness: literal Google launch figure label; benchmark-name conflict with full card/report; this cell is not used as ordinary MMLU-Pro or MMMLU. https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/

Independent evaluators

Evaluator harnesses are distinct from vendor measurements. Missing coverage is not a failed test.

Independent evaluator headline scores; expand evidence for every record
Benchmark / evaluator	Headline score	Evidence
GDPval-AA v2.1 Elo · Artificial Analysis	-36unknown effort	All 1 recorded result & sources -36 · raw -36 Elo Headline · unknown effort · Independent evaluator Source/record date: 2026-10-20 Manually observed public GDPval-AA v2.1 leaderboard table, full 298-configuration pool checked 20 October 2026. Exact source configuration: Gemma 4 E4B (Reasoning); creator: Google; source rank 277; source release-month label Apr 2026 is model release metadata, not execution date. Published rounded native Elo −36; literal displayed CI -26 / +26. Native Elo is not a percentage and is not the normalized (Elo−500)/2000 chart quantity. This versioned metric is separate from older generic GDPval-AA Elo and vendor GDPval/GDPval-AA results; no old evidence is overwritten. Execution date, inference precision, unprinted effort budgets and task costs are unpublished in this captured table. Elo is relative to the dated comparison pool; drift from an earlier observation does not prove a new evaluation run. Source says Reasoning (on), not an API effort level; reasoning effort remains unknown. Headline: highest published native Elo for this exact approved model; tied values prefer existing featured effort then source order. https://artificialanalysis.ai/evaluations/gdpval-aa
Intelligence Index · Artificial Analysis	8.9unknown effort	All 1 recorded result & sources 8.9 · raw 8.90908441839731 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA-Omniscience Index · Artificial Analysis	-19.7unknown effort	All 2 recorded results & sources -19.7 · raw -19.7 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b -41 · raw -40.983333333333334 index Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
AA-Omniscience accuracy · Artificial Analysis	8.6%unknown effort	All 2 recorded results & sources 8.6% · raw 8.583333333333334 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 8.3% · raw 8.316666666666666 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
AA-Omniscience hallucination rate · Artificial Analysis	30.9%unknown effort	All 2 recorded results & sources 30.9% · raw 30.938924339106656 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 53.8% · raw 53.772041447009634 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
GDPval-AA Elo · Artificial Analysis	-36.4unknown effort	All 1 recorded result & sources -36.4 · raw -36.44 Elo Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
Terminal-Bench 4.0 (AA) · Artificial Analysis	0%unknown effort	All 1 recorded result & sources 0% · raw 0 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
Terminal-Bench 2.1 (AA) · Artificial Analysis	1.9%unknown effort	All 1 recorded result & sources 1.9% · raw 1.8726591760299598 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
Terminal Bench Hard (AA) · Artificial Analysis	8.3%unknown effort	All 2 recorded results & sources 8.3% · raw 8.33333333333333 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 7.6% · raw 7.57575757575758 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
SciCode (AA) · Artificial Analysis	24.4%unknown effort	All 1 recorded result & sources 24.4% · raw 24.421296296296298 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA-LCR · Artificial Analysis	32%unknown effort	All 2 recorded results & sources 32% · raw 32 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 24% · raw 24 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
IFBench (AA) · Artificial Analysis	44.2%unknown effort	All 2 recorded results & sources 44.2% · raw 44.2176870748299 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 40.5% · raw 40.5442176870748 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
Humanity's Last Exam (AA harness) · Artificial Analysis	3.8%unknown effort	All 2 recorded results & sources 3.8% · raw 3.75347544022243 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 4.8% · raw 4.81927710843374 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
GPQA Diamond (AA) · Artificial Analysis	57.6%unknown effort	All 2 recorded results & sources 57.6% · raw 57.5757575757576 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 54.9% · raw 54.949494949494905 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
CritPt (AA) · Artificial Analysis	0.6%unknown effort	All 2 recorded results & sources 0.6% · raw 0.571428571428571 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 0.3% · raw 0.28571428571428603 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
GDP.PDF (AA) · Artificial Analysis	1%unknown effort	All 1 recorded result & sources 1% · raw 1 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AutomationBench (AA) · Artificial Analysis	0.5%unknown effort	All 1 recorded result & sources 0.5% · raw 0.46154542449164326 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
MMMU-Pro (AA) · Artificial Analysis	51.4%unknown effort	All 2 recorded results & sources 51.4% · raw 51.38728323699419 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 51.2% · raw 51.2138728323699 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
τ²-Telecom (AA) · Artificial Analysis	20.8%unknown effort	All 2 recorded results & sources 20.8% · raw 20.7602339181287 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 26% · raw 26.023391812865498 % Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
AA Industry Index · Finance and accounting · Artificial Analysis	7.6unknown effort	All 1 recorded result & sources 7.6 · raw 7.62938839258012 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA Industry Index · Strategy and operations · Artificial Analysis	3.2unknown effort	All 1 recorded result & sources 3.2 · raw 3.21440522005084 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA Industry Index · Legal · Artificial Analysis	11.4unknown effort	All 1 recorded result & sources 11.4 · raw 11.4082208842173 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA Industry Index · Engineering · Artificial Analysis	14.8unknown effort	All 1 recorded result & sources 14.8 · raw 14.7887356017476 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA Industry Index · Economics · Artificial Analysis	17.9unknown effort	All 1 recorded result & sources 17.9 · raw 17.9287164040778 index Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA-Briefcase Elo · Artificial Analysis	0unknown effort	All 1 recorded result & sources 0 · raw 0 Elo Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA-Briefcase analytical quality Elo · Artificial Analysis	0unknown effort	All 1 recorded result & sources 0 · raw 0 Elo Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA-Briefcase presentation Elo · Artificial Analysis	0unknown effort	All 1 recorded result & sources 0 · raw 0 Elo Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
AA-Briefcase rubric pass rate · Artificial Analysis	1.9%unknown effort	All 1 recorded result & sources 1.9% · raw 1.9191919191919191 % Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b
Output speed · Artificial Analysis	37.2 tok/sunknown effort	All 2 recorded results & sources 37.2 tok/s · raw 37.2440498041354 tok/s Headline · unknown effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Reasoning enabled without a published common API effort label. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b 41.2 tok/s · raw 41.2254665085022 tok/s Alternative · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning
Intelligence Index (estimated) · Artificial Analysis	7.5none effort	All 1 recorded result & sources 7.5 · raw 7.45765296912625 index Headline · none effort · Independent evaluator Source/record date: 2026-10-19 Independent AA exact model page explicitly links https://huggingface.co/google/gemma-4-E4B-it. Explicit non-reasoning configuration. Index v4.3.2; inference precision, execution date and exact budget unpublished. Linking reference weights does not prove BF16 inference. Costs with unavailable/zero model prices are not imported as free tariffs. https://artificialanalysis.ai/models/gemma-4-e4b-non-reasoning

Read how we select and source scores or the comparison guide.