Gemma 3 27B
google/gemma-3-27b-it
Gemma 3 27B is a 27‑billion‑parameter model from Google that accepts text and image inputs and generates text output. It provides a 262,144‑token context window, supports tools, does not support reasoning, and has open weights. The model was released on 2025‑03‑12, with pricing of $0.10 per M input tokens and $0.30 per M output tokens.
- tool use
- structured outputs
- open-weight
At a glance
Specifications
What the API gives you
- Context window
- 131K tokens
- Max output
- 118K tokens
- Input price
- $0.080 / 1M tokens
- Output price
- $0.45 / 1M tokens
- Cache read
- $0.040 / 1M tokens
- Modalities
- text, image → text
- Knowledge cutoff
- Aug 31, 2024
- Released
- Mar 12, 2025
Measured evidence
Public benchmark record
- Aider Polyglot
- 4.9%2026-07-17
- Aider Polyglot
- 4.9%2026-07-21
- Aider Polyglot
- 4.9%2026-07-23
- Aider Polyglot
- 4.9%2026-07-28
- Aider Polyglot
- 4.9%2026-07-27
- Aider Polyglot
- 4.9%2026-07-29
- Aider Polyglot
- 4.9%2026-07-30
- Aider Polyglot
- 4.9%2026-07-31
- Aider Polyglot
- 4.9%2026-08-22
- Aider Polyglot
- 4.9%2026-08-01
- Aider Polyglot
- 4.9%2026-08-02
- Aider Polyglot
- 4.9%2026-08-05
- Aider Polyglot
- 4.9%2026-08-03
- Aider Polyglot
- 4.9%2026-08-09
- Aider Polyglot
- 4.9%2026-08-04
- Aider Polyglot
- 4.9%2026-08-06
- Aider Polyglot
- 4.9%2026-08-07
- Aider Polyglot
- 4.9%2026-08-14
- Aider Polyglot
- 4.9%2026-08-08
- Aider Polyglot
- 4.9%2026-08-10
- Aider Polyglot
- 4.9%2026-08-11
- Aider Polyglot
- 4.9%2026-08-18
- Aider Polyglot
- 4.9%2026-08-12
- Aider Polyglot
- 4.9%2026-08-19
- Aider Polyglot
- 4.9%2026-08-13
- Aider Polyglot
- 4.9%2026-08-15
- Aider Polyglot
- 4.9%2026-08-20
- Aider Polyglot
- 4.9%2026-08-16
- Aider Polyglot
- 4.9%2026-08-17
- Aider Polyglot
- 4.9%2026-08-21
- Aider Polyglot
- 4.9%2026-08-27
- Aider Polyglot
- 4.9%2026-08-23
- Aider Polyglot
- 4.9%2026-08-24
- Aider Polyglot
- 4.9%2026-08-25
- Aider Polyglot
- 4.9%2026-08-26
- Aider Polyglot
- 4.9%2026-08-28
- Aider Polyglot
- 4.9%2026-08-29
- Aider Polyglot
- 4.9%2026-08-30
- Aider Polyglot
- 4.9%2026-08-31
- Aider Polyglot
- 4.9%2026-09-01
- Aider Polyglot
- 4.9%2026-09-02
- Aider Polyglot
- 4.9%2026-09-03
- Aider Polyglot
- 4.9%2026-09-04
- BFCL
- 29.5%2026-07-28
- BFCL
- 29.5%2026-07-27
- BFCL
- 29.5%2026-07-29
- BFCL
- 29.5%2026-07-30
- BFCL
- 29.5%2026-07-31
- BFCL
- 29.5%2026-08-02
- BFCL
- 29.5%2026-08-01
- BFCL
- 29.5%2026-08-05
- BFCL
- 29.5%2026-08-03
- BFCL
- 29.5%2026-08-09
- BFCL
- 29.5%2026-08-04
- BFCL
- 29.5%2026-08-06
- BFCL
- 29.5%2026-08-07
- BFCL
- 29.5%2026-08-10
- BFCL
- 29.5%2026-08-14
- BFCL
- 29.5%2026-08-08
- BFCL
- 29.5%2026-08-11
- BFCL
- 29.5%2026-08-18
- BFCL
- 29.5%2026-08-15
- BFCL
- 29.5%2026-08-12
- BFCL
- 29.5%2026-08-13
- BFCL
- 29.5%2026-08-19
- BFCL
- 29.5%2026-08-16
- BFCL
- 29.5%2026-08-20
- BFCL
- 29.5%2026-08-27
- BFCL
- 29.5%2026-08-17
- BFCL
- 29.5%2026-08-21
- BFCL
- 29.5%2026-08-22
- BFCL
- 29.5%2026-08-23
- BFCL
- 29.5%2026-08-24
- BFCL
- 29.5%2026-08-28
- BFCL
- 29.5%2026-08-25
- BFCL
- 29.5%2026-08-26
- BFCL
- 29.5%2026-08-29
- BFCL
- 29.5%2026-08-30
- BFCL
- 29.5%2026-08-31
- BFCL
- 29.5%2026-09-01
- BFCL
- 29.5%2026-09-02
- BFCL
- 29.5%2026-09-03
- BFCL
- 29.5%2026-09-04
- LMArena Creative Writing
- 1345 Elo2026-07-21
- LMArena Creative Writing
- 1344 Elo2026-07-27
- LMArena Creative Writing
- 1344 Elo2026-07-30
- LMArena Creative Writing
- 1345 Elo2026-08-02
- LMArena Creative Writing
- 1345 Elo2026-08-03
- LMArena Creative Writing
- 1345 Elo2026-08-06
- LMArena Creative Writing
- 1345 Elo2026-08-12
- LMArena Creative Writing
- 1344 Elo2026-08-21
- LMArena Creative Writing
- 1345 Elo2026-08-27
- LMArena Creative Writing
- 1345 Elo2026-09-01
- LMArena Instruction Following
- 1322 Elo2026-07-21
- LMArena Instruction Following
- 1322 Elo2026-07-27
- LMArena Instruction Following
- 1322 Elo2026-07-30
- LMArena Instruction Following
- 1322 Elo2026-08-02
- LMArena Instruction Following
- 1322 Elo2026-08-03
- LMArena Instruction Following
- 1322 Elo2026-08-06
- LMArena Instruction Following
- 1321 Elo2026-08-12
- LMArena Instruction Following
- 1321 Elo2026-08-21
- LMArena Instruction Following
- 1321 Elo2026-08-27
- LMArena Instruction Following
- 1321 Elo2026-09-01
- LMArena Long Query
- 1333 Elo2026-07-21
- LMArena Long Query
- 1333 Elo2026-07-27
- LMArena Long Query
- 1333 Elo2026-07-30
- LMArena Long Query
- 1333 Elo2026-08-02
- LMArena Long Query
- 1333 Elo2026-08-03
- LMArena Long Query
- 1333 Elo2026-08-06
- LMArena Long Query
- 1333 Elo2026-08-12
- LMArena Long Query
- 1333 Elo2026-08-21
- LMArena Long Query
- 1333 Elo2026-08-27
- LMArena Long Query
- 1333 Elo2026-09-01
- LMArena Math
- 1311 Elo2026-07-21
- LMArena Math
- 1311 Elo2026-07-27
- LMArena Math
- 1311 Elo2026-07-30
- LMArena Math
- 1311 Elo2026-08-02
- LMArena Math
- 1311 Elo2026-08-03
- LMArena Math
- 1311 Elo2026-08-06
- LMArena Math
- 1312 Elo2026-08-12
- LMArena Math
- 1312 Elo2026-08-21
- LMArena Math
- 1311 Elo2026-08-27
- LMArena Math
- 1311 Elo2026-09-01
- LMArena Text
- 1366 Elo2026-07-21
- LMArena Text
- 1366 Elo2026-07-27
- LMArena Text
- 1366 Elo2026-07-30
- LMArena Text
- 1366 Elo2026-08-02
- LMArena Text
- 1366 Elo2026-08-10
- LMArena Text
- 1366 Elo2026-08-06
- LMArena Text
- 1366 Elo2026-08-03
- LMArena Text
- 1366 Elo2026-08-21
- LMArena Text
- 1366 Elo2026-08-12
- LMArena Text
- 1365 Elo2026-08-27
- LMArena Text
- 1365 Elo2026-09-01
- LMArena Vision
- 1165 Elo2026-07-12
- LMArena Vision
- 1159 Elo2026-07-21
- LMArena Vision
- 1159 Elo2026-07-26
- LMArena Vision
- 1159 Elo2026-07-30
- LMArena Vision
- 1159 Elo2026-08-02
- LMArena Vision
- 1160 Elo2026-08-06
- LMArena Vision
- 1160 Elo2026-08-21
- LMArena Vision
- 1159 Elo2026-08-27
- SWE-rebench‡
- 5.4% ± 0.42025-09-01text
LMArena data licensed CC-BY-4.0. Aider (Apache-2.0) and SWE-bench (MIT) are open. Design Arena uses blind human preference. ‡ Submitted scaffold plus model result. Compare the harness and effort settings, not only the model name.
Open-weight model on Hugging Face.
Task evidence
How broad is the support?
Each status is exact to this model version and the selected task. Official release claims remain provisional.
- Best LLM for Debugging & Code ReviewEstablished evidence
60% of intended weighted task evidence across 2 independent sources. 6 of 6 configured evaluation feeds currently have results.
- Best LLM for Business WritingProvisional evidence
70% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Marketing CopyProvisional evidence
70% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Chat & RoleplayProvisional evidence
70% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Roleplay & CharacterProvisional evidence
70% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Tool & Function CallingProvisional evidence
65% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Agent or LMArena Search. 4 of 5 configured evaluation feeds currently have results.
- Best LLM for Creative WritingProvisional evidence
60% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for WritingProvisional evidence
60% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 5 of 5 configured evaluation feeds currently have results.
- Best Local LLM for CodingProvisional evidence
56% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LiveBench Coding or LMArena WebDev. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Hermes AgentProvisional evidence
55% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LMArena Agent or Terminal-Bench 2.1. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for OpenHandsProvisional evidence
55% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LMArena Agent or Terminal-Bench 2.1. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for GooseProvisional evidence
55% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LMArena Agent or Terminal-Bench 2.1. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for OpenClawProvisional evidence
55% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LMArena Agent or Terminal-Bench 2.1. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Image UnderstandingProvisional evidence
55% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Reasoning or VLMEvalKit tasks. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for a General AssistantProvisional evidence
55% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Reasoning. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Logic & Problem SolvingProvisional evidence
50% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Reasoning. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for AI AgentsProvisional evidence
50% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LiveBench Agentic Coding or LMArena Agent. 7 of 7 configured evaluation feeds currently have results.
- Best LLM for Competition MathProvisional evidence
40% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Mathematics or LiveBench Reasoning. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for Math & ReasoningProvisional evidence
40% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Mathematics or LiveBench Reasoning. 5 of 5 configured evaluation feeds currently have results.
- Best LLM for CodingProvisional evidence
36% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from Design Arena Coding or LiveBench Coding. 8 of 8 configured evaluation feeds currently have results.
- Best Value LLMProvisional evidence
36% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or Route reliability. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for JSON & Schema OutputProvisional evidence
33% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LiveBench Data Analysis or Structured-output eval. 4 of 5 configured evaluation feeds currently have results.
- Best LLM for Structured Data ExtractionProvisional evidence
33% of intended weighted task evidence across 2 independent sources; broader independent confirmation is still missing from LiveBench Data Analysis or Structured-output eval. 4 of 5 configured evaluation feeds currently have results.
- Best Open-Weight LLM for Self-HostingProvisional evidence
33% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or Route reliability. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for Code CompletionProvisional evidence
31% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Coding or LMArena WebDev. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for Autonomous AgentsProvisional evidence
29% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from Frontier-Bench or LiveBench Agentic Coding. 7 of 7 configured evaluation feeds currently have results.
- Best LLM for Fast SummariesProvisional evidence
29% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or Route reliability. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for Literary TranslationProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 4 of 5 configured evaluation feeds currently have results.
- Best Cheap LLM for High VolumeProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or Route reliability. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for Everyday TranslationProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 4 of 5 configured evaluation feeds currently have results.
- Cheapest Capable LLMProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or Route reliability. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for TranslationProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LiveBench Instruction Following or LiveBench Language. 4 of 5 configured evaluation feeds currently have results.
- Best LLM for Massive ContextProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or LongBench v2. 3 of 4 configured evaluation feeds currently have results.
- Best LLM for OCR & DocumentsProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for Vision & DocumentsProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for Charts & DiagramsProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for Document ParsingProvisional evidence
25% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Document or Structured-output eval. 3 of 5 configured evaluation feeds currently have results.
- Best LLM for Agentic CodingProvisional evidence
23% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from Design Arena Full-stack or Design Arena Web-apps. 8 of 8 configured evaluation feeds currently have results.
- Best LLM for Document SummarizationProvisional evidence
20% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from FACTS Grounding or LiveBench Instruction Following. 4 of 6 configured evaluation feeds currently have results.
- Best LLM for SummarizationProvisional evidence
20% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from FACTS Grounding or LiveBench Instruction Following. 4 of 6 configured evaluation feeds currently have results.
- Best LLM for Computer & Browser UseProvisional evidence
20% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from LMArena Agent or OSWorld Verified. 4 of 4 configured evaluation feeds currently have results.
- Best LLM for RAG & Long ContextProvisional evidence
18% of intended weighted task evidence across 1 independent source; broader independent confirmation is still missing from FACTS Grounding or LMArena Document. 4 of 6 configured evaluation feeds currently have results.
- Best LLM for High-Volume ExtractionProvisional evidence
13% of intended weighted task evidence across 0 independent sources; broader independent confirmation is still missing from LiveBench Data Analysis or Route reliability. 2 of 4 configured evaluation feeds currently have results.
- Best LLM for Grounded RAGProvisional evidence
10% of intended weighted task evidence across 0 independent sources; broader independent confirmation is still missing from CRAG or FACTS Grounding. 2 of 5 configured evaluation feeds currently have results.
Recommendation coverage
Where this model appears
Not currently included in a published recommendation. The task evidence panel above says exactly what is confirmed and what is still missing.
Catalog data via OpenRouter and models.dev. Last checked within the past 6 hours.