The Known Good Updated 25 Jul 2026
OpenAI ·Open weights (unrestricted) ·Released August 2025·Tracked since 25 Jul 2026

OpenAI: gpt-oss-120b — quality, speed & price analysis

Compare with another model Provider benchmarks Data sources
Quality
#33 of 61
61based on 3 of 6 evaluations
Known Good Index
Speed
#10 of 316
172.2
Output tokens per second
Price
#33 of 340
$0.07
Blended per 1M tokens
In $0.04 · Out $0.17
Cache hit price
#25 of 206
$0.015
per 1M tokens (−59%)
Context
#238 of 345
131k
tokens (~200 pages)
Summary

OpenAI: gpt-oss-120b is in the third quarter of the models we track for quality, and inexpensive for its standing, and notably fast. It accepts text input, outputs text, and supports a 131k-token context window.

It scores 61 on the Known Good Index, in line with the tracked average of 60, from 3 of the six component evaluations. Measured against the tracked average, its strongest result is MATH / AIME (89 against 53) and its weakest is Terminal-Bench (19 against 45).

Pricing is $0.04 per 1M input tokens (below the $2.18 average) and $0.17 per 1M output (below the $9.91 average). Cached input reads at $0.015 per 1M, a 59% discount on its input price. At 172 tokens per second it sits in the fastest quarter of the models we track.

Generated from tracked data on 25 Jul 2026 · how these comparisons are computed →

Technical specifications
ReasoningYes
Input modalitytext
Output modalitytext
Context window131,072 tokens
Max output131,072 tokens
WeightsOpen weights · Open weights (unrestricted)
Knowledge cutoffJune 2024
Tool callingYes
Providers17 hosts · compare →
Highlights — where OpenAI: gpt-oss-120b sits
Quality

Known Good Index · Higher is better · Composite of six published evaluations

98
96
95
94
93
93
93
91
90
90
88
88
83
61
gpt-5.5-pre-release
Gemini 3.1 Pro
Preview
Gemini 3.5 Flash
Claude Opus 4.7
Qwen3.7 Max
DeepSeek V4 Pro
Kimi K2.6
GLM 5.2
Qwen3.6 Max
Preview
GPT-5.4
Claude Opus 4.6
GLM 5.1
Gemini 3 Flash
Preview
gpt-oss-120b
Speed

Output tokens per second · Higher is better · OpenRouter endpoint telemetry

3,894
1,728
1,134
578
449
387
320
179
173
172
168
166
Relace Apply 3
Morph V3 Large
Morph V3 Fast
Grok 4.20
Multi-Agent
gpt-oss-safeguard-20b
Nano Banana 2 Lite
(Gemini 3.1 Flash Lite Image)
Mercury 2
Nano Banana
(Gemini 2.5 Flash Image)
Llama 3.2 3B
Instruct
gpt-oss-120b
Laguna XS 2.1
Nano Banana 2
(Gemini 3.1 Flash Image Preview)
Price

Blended USD per 1M tokens · Lower is better · Published pricing via OpenRouter

$0.01
$0.02
$0.04
$0.04
$0.04
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.07
Ling-2.6-flash
Mistral Nemo
Granite 4.0 Micro
Llama 3 8B Lunaris
Nex-N2-Mini
Qwen2.5 7B
Instruct
gpt-oss-20b
Mistral Small 3
Llama 3.1 8B
Instruct
MythoMax 13B
Nova Micro 1.0
Granite 4.1 8B
Gemma 3 4B
gpt-oss-120b
Jump to section

Quality

How OpenAI: gpt-oss-120b compares against every tracked model

Known Good IndexCoding IndexCapability indicesBy open weights
Known Good Index

Composite of GPQA Diamond, MATH / AIME, SWE-bench Verified, LiveBench, Humanity's Last Exam and Terminal-Bench, normalised 0–100. Scores are ingested from Epoch AI; we do not run these evaluations.

14 of 908 models ▾ ☰⚙⇩ +Add model from specific provider
98
96
95
94
93
93
93
91
90
90
88
88
83
61
💡 Reasoning models are indicated by a lightbulb
The Known Good

Quality Breakdown

OpenAI: gpt-oss-120b on each evaluation feeding the index, against the tracked average

Evaluation scores

Bar = OpenAI: gpt-oss-120b · grey marker = average across tracked models · scores ingested from Epoch AI

GPQA Diamond76 / 63
MATH / AIME89 / 53
Terminal-Bench19 / 45
The Known Good

Arena Elo

Human preference rating from blind pairwise votes

Text Arena

Elo from blind pairwise votes, ingested from Arena

1366
Elo · 95% CI ±4
#157
of 381 ranked models
31k
votes counted
1352
style-controlled (#181)

Ingested 25 Jul 2026.

The Known Good

Price & Cost

Published API pricing by token type

BlendedInput vs outputCache discountQuality per dollar
Blended Price

USD per 1M tokens at a 3:1 input:output blend. Live from OpenRouter, cross-checked against models.dev. Lower is better.

14 of 908 models ▾ ☰⚙⇩ +Add model from specific provider
$0.01
$0.02
$0.04
$0.04
$0.04
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.06
$0.07
The Known Good
We publish blended price rather than cost-per-task. Cost-per-task requires token counts from an evaluation suite we do not run. Blended price is computed from published rates and is fully reproducible — see the pricing methodology.

Context Window

Maximum input tokens accepted

Context windowMax output tokens
Context Window

Maximum input context from provider metadata via models.dev and OpenRouter. Higher is better.

13 of 908 models ▾ ☰⚙⇩ +Add model from specific provider
2.00M
2.00M
2.00M
2.00M
2.00M
1.31M
1.05M
1.05M
1.05M
1.05M
1.05M
1.05M
131k
The Known Good

Speed

Median output throughput across hosting providers

Output speedBy provider
Output Speed

Median output tokens per second across all tracked providers. Derived from OpenRouter endpoint telemetry. Higher is better.

12 of 908 models ▾ ☰⚙⇩ +Add model from specific provider
3,894
1,728
1,134
578
449
387
320
179
173
172
168
166
The Known Good

Latency

Time to first token, median across providers

Time to first tokenEnd-to-end response
Latency (TTFT)

Seconds to first streamed chunk, median across tracked providers. Derived from OpenRouter endpoint telemetry. Lower is better.

13 of 908 models ▾ ☰⚙⇩ +Add model from specific provider
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.3
0.3
0.7
The Known Good

Providers

Hosts serving OpenAI: gpt-oss-120b, with their own price and speed

17 providers
Full provider benchmarks →
ProviderInput $/MOutput $/M BlendedSpeedTTFTContext
Cerebras $0.35 $0.75 $0.45 495 0.2s 131k
Groq $0.15 $0.60 $0.26 411 0.1s 131k
SambaNova $0.14 $0.95 $0.34 306 0.6s 131k
BaseTen $0.10 $0.50 $0.20 285 0.2s 128k
Nebius $0.15 $0.60 $0.26 281 0.2s 131k
DeepInfra $0.04 $0.17 $0.07 246 0.4s 131k
Parasail $0.10 $0.75 $0.26 207 0.3s 131k
Mara $0.15 $0.75 $0.30 181 1.4s 131k
Amazon Bedrock · first party $0.15 $0.60 $0.26 166 0.4s 131k
Together $0.15 $0.60 $0.26 111 0.2s 131k
Google · first party $0.09 $0.36 $0.16 85 0.4s 131k
Phala $0.15 $0.60 $0.26 82 0.8s 131k
DigitalOcean $0.07 $0.49 $0.17 62 0.5s 128k
Novita $0.05 $0.25 $0.10 42 2.7s 131k
CoreWeave $0.03 $0.17 $0.07 41 0.4s 131k
Mancer 2 $0.06 $0.50 $0.17 41 1.0s 131k
SiliconFlow $0.05 $0.45 $0.15 26 1.3s 131k

Per-host pricing and telemetry from OpenRouter. Speed is output tokens per second; TTFT is time to first token.

The Known Good

History

How price and quality have moved since we began tracking this model

Quality & price over time

Since 25 Jul 2026

Known Good Index 61 · based on 3 of 6 evaluations
Blended price $0.07 per 1M · observed 25 Jul 2026
Output speed 172 tok/s · observed 25 Jul 2026
Time to first token 0.7s

No recorded changes for this model since tracking began.

The Known Good

Comparisons

Head-to-head pages pairing OpenAI: gpt-oss-120b with other tracked models

OpenAI: gpt-oss-120b vs gemini-2.0-flash-thinking-exp-01-21 OpenAI: gpt-oss-120b vs OpenAI: GPT-5 Mini OpenAI: gpt-oss-120b vs claude-opus-4-20250514 OpenAI: gpt-oss-120b vs DeepSeek: DeepSeek V3 0324 OpenAI: gpt-oss-120b vs OpenAI: o1 OpenAI: gpt-oss-120b vs claude-sonnet-4-5-20250929 OpenAI: gpt-oss-120b vs Google: Gemini 2.5 Pro OpenAI: gpt-oss-120b vs OpenAI: GPT-5 Nano All 49 comparisons →
Sources for this page
Pricing & providers — OpenRouter Metadata — models.dev Benchmarks — Epoch AI (CC-BY 4.0)Elo — Arena Full attribution →