Skip to content
Model Pareto

gpt-oss-120b

OpenAIOpen weightsopenai.com ↗
Compare with nearest rivalsCompare with…
Input
$0.15/M tok
Output
$0.595/M tok
Blended
$0.261/M tok · 3:1
Speed
186tok/s
Context
131Ktokens

Open weights; price is the median third-party host price (Artificial Analysis). Weights ship in MXFP4 (card: "post-trained with MXFP4 quantization of the MoE weights, making gpt-oss-120b run on a single 80GB GPU").

Rankings

Self-hostingEstimated

How we estimate →

What it would cost to run these open weights yourself on rented GPUs, compared with the API price above.

Hardware
2× H200 141GB
FP8 weights
Throughput
~10,270 tok/s
many requests batched
Price
$0.076–0.62 /M tok
busy → light use
On your own machine
Mac Studio M5 Max 128GB (int4, ~170 tok/s single-stream)
single consumer GPU or Mac
Assumptions (5)
  • FP8 weights (117B params, 5.1B active per token) + 25% KV-cache headroom ≈ 146 GB
  • 2× H200 141GB at $3.59–$7.91/GPU-hour on-demand (2026-09-24)
  • ~10,270 output tok/s aggregate at batch 369 (bandwidth-bound); MoE compute scales with active params
  • Blended 3:1 input:output; prefill ~183,349 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 Pro (Public)Estimated
~#18 of 19 (estimated)~46.9%
  1. #1Claude Opus 5.589.9%
  2. #2Claude Sonnet 5.581.3%
  3. #17Muse Glimmer51.2%
  4. ~#18gpt-oss-120bEstimated46.9%
  5. #18Nex-N2.5-Mini43.8%
No published score — estimated from related benchmarks
DeepSWE v1.1Estimated
~#28 of 31 (estimated)~43.3%
  1. #1DeepSeek V4.1 Flash74.2%
  2. #2Grok 4.772.6%
  3. #27Qwen3.8 Max (0902)51.0%
  4. ~#28gpt-oss-120bEstimated43.3%
  5. #28Qwen3.8 27B42.2%
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. ~#15gpt-oss-120bEstimated<41.1
No published score — estimated from related benchmarks

Writing & chat

LMArena Text, Non-English (Elo)Verified
#43 of 531338 Elo
  1. #1Claude Opus 5.51501
  2. #2Claude Fable 5.1Few votes1496
  3. #42Trinity Large Thinking1353
  4. #43gpt-oss-120b1338
  5. #44INTELLECT-3Few votes1327
lmarena.ai · 2026-09-24
LMArena Text (Elo)Verified
#44 of 531352 Elo
  1. #1Claude Opus 5.51509
  2. #2Claude Fable 5.11501
  3. #43INTELLECT-31356
  4. #44gpt-oss-120b1352
  5. #45Llama 3.1 Nemotron Ultra 253B v1Few votes1348
lmarena.ai · 2026-09-24

Hard reasoning

Agents

OSWorld-VerifiedEstimated
~#11 of 12 (estimated)~63.5%
  1. #1Kimi K384.8%
  2. #2Qwen3.8 27B84.3%
  3. #10Muse Glimmer65.9%
  4. ~#11gpt-oss-120bEstimated63.5%
  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. ~#5gpt-oss-120bEstimated<47.5%
No published score — estimated from related benchmarks
BrowseCompEstimated
~#15 of 18 (estimated)~57.1%
  1. #1Atria Dawn Preview92.5%
  2. #2GPT-6 Astra91.5%
  3. #14Inkling77.1%
  4. ~#15gpt-oss-120bEstimated57.1%
  5. #15Mistral Medium 3.548.6%
No published score — estimated from related benchmarks