Foundation Models benchmarks
How performance is measured at the frontier · as of 2026-11-15 · 7 benchmarks
Capability
LMArena Text Arena
Arena ScoreStyle-controlled, crowd-voted head-to-head preference rating.
↑ higher is better·LMArena
- #011,526
- #021,522
- #031,511
- #041,507
- #051,504
- #061,503
- #071,502
- #081,501
- #091,497
- #101,497
- #111,496
- #121,494
LiveBench
% scoreObjective, contamination-resistant evaluation across reasoning, coding, mathematics, data analysis, language, and instruction following.
↑ higher is better·LiveBench
- #0188.0%
- #0286.5%
- #0386.1%
- #0485.4%
- #0584.1%
- #0683.6%
- #0782.5%
- #0882.1%
- #0981.9%
- #1079.2%
- #1178.1%
- #1278.0%
SWE-bench Verified
% resolvedShare of real GitHub issues the model resolves end-to-end.
↑ higher is better·SWE-bench
- #0171.8%
- #02
Moatless Tools + Claude 4 Sonnet 70.8% - #03
OpenHands + Claude 4 Sonnet 70.4% - #04
SWE-agent + Claude 4 Sonnet 66.6% - #0565.8%
- #0665.4%
- #07
SWE-agent + Claude 3.7 Sonnet w/ Review Heavy 62.4% - #08
OpenHands + 4x Scaled (2024-02-03) 60.8% - #0958.2%
- #10
OpenHands + CodeAct v2.1 (claude-3-5-sonnet-20241022) 53.0% - #11
Composio SWE-Kit (2024-10-25) 48.6% - #1247.2%
Agents
LMArena Agent Arena
IPS scoreReal-session agent quality across task outcomes, steerability, recovery, and tool use.
↑ higher is better·LMArena Agent Arena
- #010.304
- #020.195
- #030.170
- #040.160
- #050.148
- #060.145
- #070.125
- #080.120
- #090.107
- #100.090
- #110.073
- #120.072
Reasoning
ARC-AGI-2
% accuracyVerified accuracy on novel abstract-reasoning tasks from ARC Prize's semi-private set.
↑ higher is better·ARC Prize
- #0195.00%
- #0294.17%
- #0394.17%
- #0493.75%
- #0593.33%
- #0693.33%
- #0792.50%
- #0892.50%
- #0992.50%
- #1092.08%
- #1192.08%
- #1292.08%
Economics
Blended cost
$ / 1M tokensPublished price, 3:1 input:output blend — lower is cheaper.
↓ lower is better·LiteLLM (open pricing)
- #01$0.45
- #02$1.71
- #03$1.98
- #04$2.15
- #05$3.00
- #06$3.38
- #07$4.50
- #08$4.50
- #09$6.00
- #10$8.00
- #11$10.00
Context
- #011M
- #021M
- #031M
- #041M
- #051M
- #061M
- #07922K
- #08922K
- #09922K
- #10500K
- #11262.1K