ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

gpt-oss-120bvsQwen3-Max

gpt-oss-120b leads 2 of 3 shared benchmarks. Qwen3-Max has the larger context window (262K tokens).

Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.

Shared benchmarks
2 – 1
gpt-oss-120b leads
Cheaper per token
list price, input + output
Larger context
Qwen3-Max
262K tokens
Provider uptime (30d)
94.2446% · —
OpenAI · Qwen
gpt-oss-120b
OpenAI · released 2025-08-05
text
Context
131K
Max out
131K
Input /1M
Output /1M
Cached /1M
Scores
6 · 5 core
Qwen3-Max
Qwen · released 2025-09-23
text
Context
262K
Max out
Input /1M
$1.2
Output /1M
$6
Cached /1M
$0.12
Scores
5 · 5 core
Quality

Benchmark matrix

Benchmarkgpt-oss-120bQwen3-MaxΔEdge
Reported by both · 3
SWE-bench Verified62.4% 69.6% -7.2 ptQwen3-Max
GPQA Diamond80.1% 72.6% epoch+7.5 ptgpt-oss-120b
OTIS Mock AIME 2024-2025 (Epoch AI run)88.9% epoch73.3% epoch+15.6 ptgpt-oss-120b
Only gpt-oss-120b reports · 2
AIME 202590% matharenanot reported
LMArena Elo1365.9 not reported
Only Qwen3-Max reports · 2
FrontierMath Tiers 1-3 v2 (Epoch AI run)not reported18.9% epoch
SimpleQA Verifiednot reported48.7% epoch
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is gpt-oss-120b minus Qwen3-Max in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing

Side by side

Specgpt-oss-120bQwen3-MaxEdge
Context window131K tokens262K tokensQwen3-Max
Max output131K tokens
Input price / 1M$1.2
Output price / 1M$6
Cached input / 1M$0.12
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$7.2
ModalitiestexttextTie
Released2025-08-052025-09-23
Cited benchmark scores65
Reliability

Provider status

All providers
More matchups

gpt-oss-120b vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run gpt-oss-120b and Qwen3-Max on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.