NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/Falcon-180B
TII logo

Falcon-180B

by TII · Released Jan 2024

Open Source
50.3
avg score
Rank #151
Compare
Better than 53% of all models
Context
N/A
Input $/1M
n/a
Output $/1M
n/a
Type
text
License
Open Source
Benchmarks
17 tested
Data as of
About

Tested on 17 benchmarks · BenchGecko score 50.3. Top scores: HellaSwag (85.3%), TriviaQA (79.9%), LAMBADA (79.8%).

Capabilities
reasoning
11.8
#173 globally
math
28.6
#205 globally
knowledge
56.3
#89 globally
language
32.6
#142 globally
general
21.9
#162 globally
Benchmark Scores
Compare All
Tested on 17 benchmarks · Ranked across 5 categories
Score Distribution (all 319 models)
0255075100
▲ You are here
reasoningCompare reasoning →
BBH

BIG-Bench Hard. 23 challenging tasks from BIG-Bench where prior language models fell below average human performance.

16.1·
MUSR

HuggingFace MuSR (Multi-Step Reasoning). Tests multi-hop reasoning requiring chaining multiple facts together.

7.5·
mathCompare math →
GSM8K

Grade school math word problems. 8,500 problems testing multi-step arithmetic reasoning. A foundational math benchmark.

54.4·
MATH Level 5

HuggingFace evaluation of MATH Level 5 problems. Competition math requiring advanced reasoning and proof construction.

2.8·
knowledgeCompare knowledge →
HellaSwag

Sentence completion requiring commonsense reasoning about physical and social situations. Tests real-world understanding.

85.3·
TriviaQA

Trivia questions sourced from trivia enthusiasts and quiz websites. Tests breadth of general knowledge.

79.9·
LAMBADA

Language modeling benchmark testing ability to predict the last word of passages requiring long-range context understanding.

79.8·
Quick compare:Google DeepMind logovs Gemma 2 27Bz-ai logovs GLM 5.3 FlashMistral AI logovs Mistral 7B V0.1
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Similar Models
Google DeepMind logo
Gemma 2 27B
Google DeepMind
50.1$0.65/1M
z-ai logo
GLM 5.3 Flash
z-ai
50.5$0.15/1M
Mistral AI logo
Mistral 7B V0.1
Mistral AI
50.0n/a
Links
Info
TIIPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@TII
Source Code
GitHubHugging Face
BenchGecko API
falcon-180b
Specifications
  • Typetext
  • ContextN/A
  • ReleasedJan 2024
  • LicenseOpen Source
  • Statusbenchmark-only
Available On
TII logoTIIn/a
Categories
reasoningmathknowledgelanguagegeneral
Learn More
transformeropen weightstokens
Share & Export
Tweet
Related Models
Google DeepMind logoGemma 2 27B
50.1
z-ai logoGLM 5.3 Flash
50.5
Mistral AI logoMistral 7B V0.1
50.0
Google DeepMind logoGemini 2.0 Pro
50.6
Falcon-180B is an open-source text AI model by TII, released in January 2024. It has an average benchmark score of 50.3.

Related Models

Gemma 2 27B · Google DeepMindGLM 5.3 Flash · z-aiMistral 7B V0.1 · Mistral AIGemini 2.0 Pro · Google DeepMindGPT-5 Pro · OpenAI

Benchmarks

HellaSwagTriviaQALAMBADAWinograndePIQA

Related Pages

TII · ProviderTII · EconomyAll ModelsCompare ModelsPricingDevelopers · API

Key facts · as of 2026-04-09

  • Falcon-180B by TII. BenchGecko score 50.3, rank 151 of 319 scored models (normalized average of public benchmark scores).
  • List price n/a input · n/a output per 1M tokens (as of 2026-04-09).

How to cite · data as of 2026-04-09

Falcon-180B · benchmarks, pricing and providers. BenchGecko, data as of 2026-04-09. https://benchgecko.ai/model/falcon-180b

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP