Compare · ModelsLive · 2 picked · head to head
Claude 2.1 vs Gemma 3 27B
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Gemma 3 27B wins on 4/4 benchmarks
Gemma 3 27B wins 4 of 4 shared benchmarks. Leads in arena · general · knowledge.
Category leads
arena·Gemma 3 27Bgeneral·Gemma 3 27Bknowledge·Gemma 3 27Bmath·Gemma 3 27B
Hype vs Reality
Attention vs performance
Claude 2.1
#293 by perf·no signal
Gemma 3 27B
#217 by perf·#4 by attention
Vendor risk
Who is behind the model
Anthropic
$965.0B·Tier 1
Google DeepMind
$4.27T·Tier 1
Head to head
4 benchmarks · 2 models
Claude 2.1Gemma 3 27B
Chatbot Arena Elo · Overall
Gemma 3 27B leads by +1342.6
Claude 2.1
22.8
Gemma 3 27B
1365.4
Dtbench
Gemma 3 27B leads by +2.6
Claude 2.1
18.3
Gemma 3 27B
20.9
GPQA diamond
Gemma 3 27B leads by +19.7
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude 2.1
10.6
Gemma 3 27B
30.3
OTIS Mock AIME 2024-2025
Gemma 3 27B leads by +20.6
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude 2.1
1.9
Gemma 3 27B
22.4
Full benchmark table
| Benchmark | Claude 2.1 | Gemma 3 27B |
|---|---|---|
Chatbot Arena Elo · Overall | 22.8 | 1365.4 |
Dtbench | 18.3 | 20.9 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 10.6 | 30.3 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 1.9 | 22.4 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| — | — | — | — | |
| $0.08 | $0.45 | 131K tokens (~66 books) | $1.72 |