The Known Good Updated 29 Jul 2026
OpenAI ·Proprietary ·Released April 2025·Tracked since 25 Jul 2026

OpenAI: o3 — quality, speed & price analysis

Compare with another model Provider benchmarks Data sources
Quality
#25 of 62
74based on 5 of 7 evaluations
Known Good Index
Speed
#85 of 317
70.8
Output tokens per second
Price
#292 of 370
$3.50
Blended per 1M tokens
In $2.00 · Out $8.00
Cache hit price
#204 of 238
$0.50
per 1M tokens (−75%)
Context
#226 of 375
200k
tokens (~300 pages)
Summary

OpenAI: o3 is above the median of the models we track for quality, and priced close to the tracked average, with mid-range throughput. It accepts image, text, file input, outputs text, and supports a 200k-token context window.

It scores 74 on the Known Good Index, above the tracked average of 64, from 5 of the seven component evaluations. Measured against the tracked average, its strongest result is MATH Level 5 (98 against 55) and its weakest is SWE-bench Verified (62 against 71).

Pricing is $2.00 per 1M input tokens (close to the $2.10 average) and $8.00 per 1M output (below the $9.62 average). Cached input reads at $0.50 per 1M, a 75% discount on its input price. At 71 tokens per second it sits in the upper middle of the tracked range.

Generated from tracked data on 29 Jul 2026 · how these comparisons are computed →

Technical specifications
ReasoningYes
Input modalityimage, text, file
Output modalitytext
Context window200,000 tokens
Max output100,000 tokens
WeightsProprietary
Knowledge cutoffJune 2024
Tool callingYes
Providers1 host · compare →
Highlights — where OpenAI: o3 sits
Quality

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

98
96
95
94
93
93
93
91
90
90
88
88
86
83
83
80
80
78
78
77
75
75
74
74
74
73
72
70
68
68
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
o3 Mini
Gemini 3 Flash
Preview
gemini-3-pro-preview
GPT-5.2
Claude Sonnet 4.6
GPT-5
Qwen3.6 Plus
GLM 5
R1
deepseek-r1
claude-3-7-sonnet-20250219
GPT-5.1
o3
gemini-2.0-pro-exp-02-05
claude-opus-4-5-20251101
o1
claude-opus-4-20250514
GLM 4.7
Speed

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

2,504
1,909
1,434
566
370
342
280
181
166
161
154
148
145
144
143
136
135
135
134
132
131
130
130
126
126
126
124
121
108
107
71
Morph V3 Large
Morph V3 Fast
Relace Apply 3
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)
Grok 4.20
Nano Banana 2
(Gemini 3.1 Flash Image Preview)
Llama 3.2 3B
Instruct
Qwen3 Next 80B A3B
Thinking
gpt-oss-120b
Gemini 3.6 Flash
Laguna XS 2.1
Muse Spark 1.1
Trinity Large
Thinking
o3 Mini High
Qwen3 30B A3B
Thinking 2507
gpt-oss-20b
Nano Banana 2
(Gemini 3.1 Flash Image)
Gemini 3.1 Flash
Lite Preview
GPT-5.6 Luna Pro
Qwen3 VL 8B
Thinking
Gemini 3.5 Flash
Llama 3.2 1B
Instruct
Gemini 3.1 Flash
Lite
Grok Build 0.1
Claude Opus 4.8
(Fast)
Nemotron 3 Nano
30B A3B
o3this model #85/317
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
$0.06
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.07
$0.08
$0.09
$0.09
$0.09
$0.10
$0.10
$0.10
$0.10
$0.11
$3.50
Ling-2.6-flash
Mistral Nemo
Granite 4.0 Micro
Llama 3 8B Lunaris
Nex-N2-Mini
Qwen2.5 7B
Instruct
gpt-oss-20b
Qwen3.7 Flash
Mistral Small 3
Llama 3.1 8B
Instruct
MythoMax 13B
Nova Micro 1.0
Granite 4.1 8B
Gemma 3 4B
Command R7B
(12-2024)
GPT-5 Nano (batch)
gpt-oss-120b
Llama 3.2 1B
Instruct
Gemma 3 12B
Laguna XS 2.1
Gemma 3n 4B
Qwen3 30B A3B
Instruct 2507
Nemotron 3 Nano
30B A3B
Gemini 2.5 Flash Lite
(batch)
Phi 4
Hy3 preview
Ministral 3 3B
2512
Reka Edge
Nova Lite 1.0
Qwen3.5-9B
o3this model #292/370 · off scale
Jump to section

Quality

How OpenAI: o3 compares against every tracked model

Known Good IndexCapability indices
Known Good Index

Composite of GPQA Diamond, Mock AIME 2024–25, MATH Level 5, 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.

⇩
98
96
95
94
93
93
93
91
90
90
88
88
86
💡 A lightbulb marks a model its provider publishes as a reasoning model
The Known Good

Quality Breakdown

OpenAI: o3 on each evaluation feeding the index, against the tracked average

Evaluation scores

Bar = OpenAI: o3 · grey marker = average across tracked models · scores ingested from Epoch AI

GPQA Diamond82 / 63
Mock AIME 2024–2584 / 53
MATH Level 598 / 55
SWE-bench Verified62 / 71
Humanity's Last Exam20 / 17
The Known Good

Arena Elo

Human preference rating from blind pairwise votes

Arena Elo

Elo from blind pairwise votes, ingested from Arena

OpenAI: o3 does not currently appear in any arena we ingest.

The Known Good

Price & Cost

Published API pricing by token type

BlendedInput vs outputQuality 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.

⇩
$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
$3.50
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.

⇩
2.00M
2.00M
2.00M
2.00M
2.00M
1.31M
1.05M
1.05M
1.05M
1.05M
1.05M
1.05M
200k
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.

⇩
2,504
1,909
1,434
566
370
342
280
181
166
161
154
148
71
The Known Good

Latency

Time to first token, median across providers

Latency (TTFT)

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

⇩
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.3
0.3
3.6
The Known Good

Providers

Hosts serving OpenAI: o3, with their own price and speed

1 provider
Full provider benchmarks →
ProviderInput $/MOutput $/M BlendedSpeedTTFTContext
OpenAI · first party $2.00 $8.00 $3.50 84 2.2s 200k

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 74 · based on 5 of 7 evaluations
Blended price $3.50 per 1M · observed 29 Jul 2026
Output speed 71 tok/s · observed 29 Jul 2026
Time to first token 3.6s

No recorded changes for this model since tracking began.

The Known Good

Comparisons

Head-to-head pages pairing OpenAI: o3 with other tracked models

OpenAI: o3 vs gemini-2.0-pro-exp-02-05 OpenAI: o3 vs OpenAI: GPT-5.1 OpenAI: o3 vs claude-3-7-sonnet-20250219 OpenAI: o3 vs DeepSeek: R1 OpenAI: o3 vs deepseek-r1 OpenAI: o3 vs claude-opus-4-5-20251101 OpenAI: o3 vs Z.ai: GLM 5 OpenAI: o3 vs OpenAI: o1 All 49 comparisons →
Sources for this page
Pricing & providers — OpenRouter Metadata — models.dev Benchmarks — Epoch AI (CC-BY 4.0) Full attribution →