Skip to content
Model Pareto

K-EXAONE 2.0 0803

LG AI Research (EXAONE)Open weightsartificialanalysis.ai ↗
Compare with nearest rivalsCompare with…
Input
—
Output
—
Speed
~82tok/sest.
Context
262Ktokens

Long-tail import from Artificial Analysis (2026-09-24); AA variant: K-EXAONE 2.0 0803. Price = AA median host price.

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs. No API sells this model, so this is its price on the chart.

Hardware
8× B200 180GB
FP8 weights
Throughput
~19,157 tok/s
many requests batched
Price
$0.29–2.58 /M tok
busy → light use
On your own machine
Mac Studio M5 Ultra 512GB (int4, ~45 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • FP8 weights (749B params, 37B active per token) + 25% KV-cache headroom ≈ 936 GB
  • 8× B200 180GB at $5.98–$14.24/GPU-hour on-demand (2026-09-24)
  • ~19,157 output tok/s aggregate at batch 548 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~229,865 tok/s
  • Low = 75% utilization at the low GPU price; high = 20% at the high price

Benchmark scores

~ italic, dashed = no published score yet, estimated from related benchmarks.

Coding

SWE-bench VerifiedEstimated
~#11 of 14 (estimated)~72.7%
  1. #1Claude Opus 596.0%
  2. #2Claude Sonnet 585.2%
  3. #10Claude Haiku 4.573.3%
  4. ~#11K-EXAONE 2.0 0803Estimated72.7%
  5. #11Nemotron 3 Ultra70.7%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#28 of 31 (estimated)~46.6%
  1. #1DeepSeek V4.1 Flash74.2%
  2. #2Grok 4.772.6%
  3. #27Qwen3.8 Max (0902)51.0%
  4. ~#28K-EXAONE 2.0 0803Estimated46.6%
  5. #28Qwen3.8 27B42.2%
No published score — estimated from related benchmarks
Terminal-Bench 4.0Estimated
~#27 of 84 (estimated)~7.9%
  1. #1Claude Sonnet 5.563.6%
  2. #2Claude Opus 5.559.6%
  3. #26Qwen3.8 2.4T A95B11.1%
  4. ~#27K-EXAONE 2.0 0803Estimated7.9%
  5. #27Qwen3.8 27B5.6%
No published score — estimated from related benchmarks
LiveCodeBenchEstimated
~#8 of 20 (estimated)~70.7%
  1. #1Qwen3.8-Omni-Flash92.6%
  2. #2Qwen3.8 Flash-Next91.9%
  3. #7gpt-oss-20b77.7%
  4. ~#8K-EXAONE 2.0 0803Estimated70.7%
  5. #8Mistral Small 464.0%
No published score — estimated from related benchmarks
AA Coding Agent IndexEstimated
~last of 15 (estimated)below measured range (<41.1)
  1. #1Claude Opus 5.566.0
  2. #2Claude Fable 5.162.2
  3. #14GPT-6 Luna41.1
  4. ~#15K-EXAONE 2.0 0803Estimated<41.1
No published score — estimated from related benchmarks

Hard reasoning

AIME (latest)Estimated
~#9 of 25 (estimated)~88.2%
  1. #1Inkling97.1%
  2. #2Muse Glimmer94.7%
  3. #8Nova 2.0 Pro Preview89.0%
  4. ~#9K-EXAONE 2.0 0803Estimated88.2%
  5. #9Mistral Medium 3.586.3%
No published score — estimated from related benchmarks

Agents

OSWorld-VerifiedEstimated
~#11 of 12 (estimated)~63.8%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #10Muse Glimmer65.9%
  4. ~#11K-EXAONE 2.0 0803Estimated63.8%
  5. #11Claude Haiku 4.550.7%
No published score — estimated from related benchmarks
OSWorld 2.0Estimated
~last of 5 (estimated)below measured range (<47.5%)
  1. #1GLM-5.3 Flash59.1%
  2. #2Kimi K358.3%
  3. #4GPT-5.547.5%
  4. ~#5K-EXAONE 2.0 0803Estimated<47.5%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#15 of 18 (estimated)~57.3%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #14Inkling77.1%
  4. ~#15K-EXAONE 2.0 0803Estimated57.3%
  5. #15Mistral Medium 3.548.6%
No published score — estimated from related benchmarks
GDPval-AA (Elo)Estimated
~#31 of 87 (estimated)~1133 Elo
  1. #1Claude Opus 5.51846
  2. #2Claude Fable 5.11735
  3. #30Ling-3.0-flash-VL1150
  4. ~#31K-EXAONE 2.0 0803Estimated1133
  5. #31Nex-N2-Pro1091
No published score — estimated from related benchmarks