ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

GLM-4.5vsgpt-oss-20b

GLM-4.5 leads 2 of 2 shared benchmarks. gpt-oss-20b has the larger context window (131K tokens).

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

Shared benchmarks
2 – 0
GLM-4.5 leads
Cheaper per token
list price, input + output
Larger context
gpt-oss-20b
131K tokens
Provider uptime (30d)
— · 94.9045%
Z.ai · OpenAI
text
Context
128K
Max out
98K
Input /1M
$0.6
Output /1M
$2.2
Cached /1M
$0.11
Scores
2 · 2 core
gpt-oss-20b
OpenAI · released 2025-08-05
text
Context
131K
Max out
131K
Input /1M
Output /1M
Cached /1M
Scores
6 · 5 core
Quality

Benchmark matrix

BenchmarkGLM-4.5gpt-oss-20bΔEdge
Reported by both · 2
AIME 202593.3% matharena89.2% matharena+4.1 ptGLM-4.5
LMArena Elo1430.2 1287.3 +142.9GLM-4.5
Only GLM-4.5 reports · 0
GLM-4.5 reports nothing gpt-oss-20b does not.
Only gpt-oss-20b reports · 3
SWE-bench Verifiednot reported60.7%
GPQA Diamondnot reported71.5%
OTIS Mock AIME 2024-2025 (Epoch AI run)not reported65.3% 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 GLM-4.5 minus gpt-oss-20b 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

SpecGLM-4.5gpt-oss-20bEdge
Context window128K tokens131K tokensgpt-oss-20b
Max output98K tokens131K tokensgpt-oss-20b
Input price / 1M$0.6
Output price / 1M$2.2
Cached input / 1M$0.11
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$2.8
ModalitiestexttextTie
Released2025-08-05
Cited benchmark scores26
Reliability

Provider status

All providers
More matchups

gpt-oss-20b vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run GLM-4.5 and gpt-oss-20b on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.