Skip to content
ModelCap

gpt-oss-20b

Current board #68Arena #238Open weightsFree tier

OpenAI·openai/gpt-oss-20b

gpt-oss-20b is a language model published by OpenAI. 12 OpenRouter endpoints are published in the latest snapshot. It accepts text input and returns text. The published context limit is 131K tokens. The catalogue declares support for reasoning, tool use, structured outputs.

ModelCap Index · Current board
41.8#68
Measured · 80% support · score interval 41.8–49.0
Market Gravity
55.2
Secondary market signal
Input
$0.03
per 1M tokens
Output
$0.13
per 1M tokens
Context
131K
tokens
Providers
12
endpoints
Gateway spend
—
latest · Vercel share

Providers

Uptime measured over the last 30 minutes
ProviderInputOutputContextQuantUptime
CoreWeavefp4Cheapest$0.03$0.13131Kfp498.88%
DeepInfrabf16$0.03$0.14131Kbf16100.0%
Parasailfp4$0.03$0.15131Kfp476.79%
Novitafp4$0.04$0.15131Kfp4100.0%
Phala$0.04$0.15131K—99.80%
Amazon Bedrockeu-west-1$0.07$0.15131K——
Amazon Bedrock$0.07$0.15131K—100.0%
SiliconFlowfp8$0.04$0.18131Kfp897.18%
Together$0.05$0.20131K——
Googleus-central1$0.07$0.25131K—57.45%
Fireworks$0.07$0.30131K——
Groq$0.075$0.30131K—99.87%

Hugging Face Inference Providers

Exact repository identity

Provider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.

ProviderInputOutputContextTTFTThroughputFeatures
deepinfra
live
$0.03$0.14131K273 ms110.2 tok/stools · structured
novita
live
$0.04$0.15131K565 ms35.0 tok/sstructured
ovhcloud
live
$0.05$0.18131K266 ms79.7 tok/stools · structured
nscale
live
$0.05$0.20131K579 ms163.0 tok/stools · structured
together
live
$0.05$0.20131K227 ms118.4 tok/stools · structured
fireworks-ai
live
$0.07$0.30131K317 ms159.4 tok/stools · structured
groq
live
$0.10$0.50131K403 ms612.7 tok/stools · structured
featherless-ai
live
——————

Reported evaluation leads (20)

Excluded from score

Structured rows reported in this exact model repository at the pinned revision below. They are useful research leads, but remain quarantined until dataset revision, harness, configuration and identity pass ModelCap’s independent admission review.

Dataset / taskMetricValueRevisionProvenance
cais/hle
hle
—8.8
Unpinned
not published
Reported
Quarantined
cais/hle
hle
—7
Unpinned
not published
Reported
Quarantined
cais/hle
hle
—6.3
Unpinned
not published
Reported
Quarantined
cais/hle
hle
—4.2
Unpinned
not published
Reported
Quarantined
cais/hle
hle
—17.3
Unpinned
not published
Reported
Quarantined
cais/hle
hle
—10.9
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—58.5859
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—67.1
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—66
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—58
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—56.8
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—74.2
Unpinned
not published
Reported
Quarantined
Idavidrein/gpqa
diamond
—71.5
Unpinned
not published
Reported
Quarantined
LEXam-Benchmark/LEXam
open_question
—32.12
Unpinned
not published
Reported
Quarantined
LEXam-Benchmark/LEXam
mcq_4_choices
—40.78
Unpinned
not published
Reported
Quarantined
LiquidAI/ifstruct-v1.0
ifstruct_v1
—91.95
Unpinned
not published
Reported
Quarantined
SWE-bench/SWE-bench_Verified
swe_bench_%_resolved
—53.2
Unpinned
not published
Reported
Quarantined
SWE-bench/SWE-bench_Verified
swe_bench_%_resolved
—37.4
Unpinned
not published
Reported
Quarantined
SWE-bench/SWE-bench_Verified
swe_bench_%_resolved
—60.7
Unpinned
not published
Reported
Quarantined
TIGER-Lab/MMLU-Pro
mmlu_pro
—73.6
Unpinned
not published
Reported
Quarantined

ModelCap Index · Current board

Methodology
41.8
Measured · 2 public benchmark observations across 2 boards
Current board
#68
Public rank basis
Measured (measured)
Index support
79.6% · strong
Index interval
41.8–49.0
Rank posterior
Not available
Top 5 / Top 10 probability
Not available / Not available
Posterior as of
Snapshot timestamp unavailable
Identity binding
Exact catalogue product · openai/openai/gpt-oss-20b · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata openai/gpt-oss-20b@6cee5e81ee83917806bbde320786a8fb61efebee (not the evaluation revision)
Direct benchmark observations
2
Breadth-adjusted BES point estimate
45.4
Bayesian Evidence Score · measured evidence input only

BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.

Observed capability
45.4
Evidence status
confirmed · 82% mass
General preference
44.4
Coding
49.4
Agents & tools
—
Reasoning
—
Evidence breadth
92%
Evidence coverage
80%
Market and catalogue signals · secondary, never the language Index rank
Usage (OpenRouter popularity)
33.6
Liquidity (providers × uptime)
86.9
Open reach (HF downloads)
91.9
Surface (context / tools / modalities)
70.2
Freshness
24.7
  • · OpenRouter popularity #44
  • · 11 live providers
  • · HF 30d downloads 8229095
  • · Catalogue: context 131072, reasoning, tools
  • · First seen 2025-08-05

These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.

BES evidence selected · measured input only

The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.

Deployment readiness

Metadata index
84.7
strong · 94% metadata coverage
Not capability
Artifact reproducibility
85.0
Access & legal clarity
100.0
Deployability
97.4
Evaluation provenance
50.0

Missing: artifactReproducibility.base-lineage, evaluationProvenance.evaluation-dataset-revisions.

A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.

Market activity

Venue coverage
Market Gravity
55.2
OpenRouter weekly popularity
#44 of 298

Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.

Arena preference

Status
Ranked
Source
Arena
Official rank
#238 of 385
Rating
1317.4
95% confidence interval
1311–1324
Votes
11K
Source snapshot
6 Aug 2026, 13:00 UTC
Category
overall
Identity match confidence
100%

Market Gravity breakdown

Usage55%
33.6

OpenRouter weekly popularity across the full model catalogue.

Liquidity25%
86.9

Independent providers versus the model's open or closed cohort, weighted by uptime.

Open reach15%
91.9

Hugging Face 30-day downloads for open models; neutral for closed models.

Freshness5%
24.7

Time since first public availability, on a six-month half-life.

Specification

Context window
131K tokens
Max output
131K tokens
Inputs
Text
Outputs
Text
Tokenizer
GPT
Cached input tokens
$0.03 / 1M
First seen on OpenRouter
5 Aug 2025
Knowledge cutoff
2024-06-30

Capabilities

  • Supported: Reasoning
  • Supported: Tool use
  • Supported: Structured output
  • Supported: Response format
  • Not supported: Moderated

Weights & access

Open weights. Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI.

Access
Open weights
Licence
apache-2.0
Pinned revision
6cee5e81ee83
Parameters
21.5B
Architecture
GptOssForCausalLM
Model type
gpt_oss
Hugging Face downloads (30d)
8.2M
Downloads (all time)
88.2M
Likes
5K
Repository
openai/gpt-oss-20b
Repository updated
26 Aug 2025, 17:25 UTC