Model
Model explorer

Falcon 2 11B

OPEN
TII · Falcon 2 family · released May 13, 2024

Second-gen Falcon trained on 5.5T tokens across 11 languages; beats Llama 3 8B, on par with Gemma 7B; also shipped with a vision-to-language (VLM) variant.

ReasoningCodingVisionFunction callingTool useAgentic
262.9
Elo · rank #353
Parameters
11B
Active params
11B (dense)
Context
8K tokens
Architecture
Causal decoder-only Transformer, 11B (60 layers, grouped-query attention (GQA, 32 query / 8 KV heads), RoPE, FlashAttention-2)
License
TII Falcon License 2.0
Languages
11+
API price (in/out)
No hosted API
Modalities
text
Benchmark results
Bar shows position within the tracked field; marker = field best
ARC-ChallengeReasoning59.7%#92
best: Llama 3.1 405B · 96.9%
ARC-EasyReasoning77.8%#30
best: Phi-3-medium (14B) · 97.7%
GSM8KMath53.8%#126
best: Llama 3.1 405B · 96.8%
HellaSwagReasoning82.9%#59
best: Claude 3 Opus · 95.4%
HumanEvalCoding29.6%#156
best: Claude Opus 4.5 · 99.4%
MMLUKnowledge58.4%#209
best: OpenAI o3 · 92.9%
TruthfulQAKnowledge52.6%#47
best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%
WinoGrandeReasoning78.3%#43
best: PaLM 2 · 90.9%
Run it locally
VRAM @ Q4
7 GB
VRAM @ FP16
22 GB
Fits on (Q4)
RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB
Throughput data unavailable.
Quantizations
GGUF
Falcon 2 family
Elo progression across releases
Falcon 2 11BMay 2024262.9
Resources
HFtiiuae/falcon-11BDOCShuggingface.co/blog/falcon2-11b
API price weights · each benchmark row carries its own source badge (see methodology)