vs
1:0
overall capability wins (2 ties across 3 scored). Falcon 180B leads overall.
- Pricing3:0
- Speed0:2
- Specs1:0
Verdict
Falcon 180B and Falcon 40B Instruct are closely matched on the overlapping metrics we have.
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Falcon 180B
- $0
- Falcon 40B Instruct
- $1
Falcon 180B is about Infinityx cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
TII
TII
Family
Falcon
Falcon
License
TII Falcon License
Apache 2.0
Open weights
Yes
Yes
Release
Sep 6, 2023
May 25, 2023
Knowledge cutoff
-
-
API / provider
Self-host / community
Together / Fireworks (ref.)
Modalities
text → text
text → text
Specs
Context window
2K
2K
Max output
2K
2K
Parameters
180B
40B
Pricing
Input $/1M
$0
$0.80
Output $/1M
$0
$0.80
Blended $/1M (3∶1)
$0
$0.80
Speed
tok/s
20
45
TTFT (s)
1.2
0.6
Reasoning
MMLU-Pro
—
—
GPQA Diamond
—
—
Humanity's Last Exam
—
—
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
—
AA-LCR v1.1
—
—
CritPt
—
—
MMMU-Pro
—
—
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
—
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
—
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
—
| Metric | Falcon 180B | Falcon 40B Instruct | Delta |
|---|---|---|---|
| Identity | |||
| Organization | TII | TII | — |
| Family | Falcon | Falcon | — |
| License | TII Falcon License | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Sep 6, 2023 | May 25, 2023 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Self-host / community | Together / Fireworks (ref.) | — |
| Modalities | text → text | text → text | — |
| Specs | |||
| Context window | 2K | 2K | tie |
| Max output | 2K | 2K | tie |
| Parameters | 180B | 40B | A +140 |
| Pricing | |||
| Input $/1M | $0 | $0.80 | A +$0.80 |
| Output $/1M | $0 | $0.80 | A +$0.80 |
| Blended $/1M (3∶1) | $0 | $0.80 | A +$0.80 |
| Speed | |||
| tok/s | 20 | 45 | B +25 |
| TTFT (s) | 1.2 | 0.6 | B +0.60 |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | — | — | — |
| Humanity's Last Exam | — | — | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | — | — |
| AA-LCR v1.1 | — | — | — |
| CritPt | — | — | — |
| MMMU-Pro | — | — | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | — | — |
Falcon 180B
- Context
- 2K / 2K out
- Parameters
- 180B
- Price
- $0 / $0
- Speed
- 20 tok/s · 1.2s TTFT
- Modalities
- text → text
- License
- TII Falcon License
UAE TII's open 180B model — historically one of the largest freely available LLMs.
Falcon 40B Instruct
- Context
- 2K / 2K out
- Parameters
- 40B
- Price
- $0.80 / $0.80
- Speed
- 45 tok/s · 0.6s TTFT
- Modalities
- text → text
- License
- Apache 2.0
TII Falcon 40B instruct — early open Apache-2.0 chat model widely hosted before Llama 2 dominated.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.