BlogLLM Benchmark Scores as Published, Side by Side
o3: Benchmark Scores as Published
Every result the catalog holds for o3: 31 tests in 12 areas, from 4 pages. 27 printed by OpenAI itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer OpenAIReleased 2025-04-16Context window 200KMax output 100KInput Text, imageOutput TextWeights ClosedData version 2026103001, as of 2026-10-30
Knowledge Work
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OpenAI economically important tasks (wins or ties against experts) | 33.5% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 69.1% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 2 more pages | Effort: high |
| Terminal-Bench (version not stated) | 30.2% | Anthropic, Claude Opus 4 and Claude Sonnet 4checked 2026-10-15; printed on 1 more page | - |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 83.3% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 2 more pages | Effort: high; Tools: no tools |
| AIME 2025 | 98.4% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 2 more pages | Effort: high; Tools: python |
| 88.9% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: no tools | |
| HMMT February 2025 | 93.3% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: python |
| FrontierMath Tier 1-3 | 15.8% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: python |
Multimodal Understanding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| CharXiv Reasoning (no tools) | 78.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| MMMU-Pro | 76.4% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| MMMU (validation) | 82.9% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 2 more pages | Effort: high |
| Video-MMMU | 83.3% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| ERQA | 64% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 14.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: no tools |
| Humanity's Last Exam (with tools) | 24.3% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: python + browser |
| FActScore (hallucination rate)lower is better | 5.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| LongFact-Concepts (hallucination rate)lower is better | 4.5% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| LongFact-Objects (hallucination rate)lower is better | 5.1% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Code Generation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Aider polyglot | 79.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| τ²-bench Telecom | 58.2% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| τ²-bench Retail | 80.2% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| τ²-bench Airline | 64.8% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| τ-bench Airline | 52% | Anthropic, Claude Opus 4 and Claude Sonnet 4checked 2026-10-15; printed on 1 more page | - |
| τ-bench Retail | 70.4% | Anthropic, Claude Opus 4 and Claude Sonnet 4checked 2026-10-15; printed on 1 more page | - |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 49.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Instruction Following
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MultiChallenge | 60.4% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| COLLIE | 98.4% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Multilingual
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MMMLU | 88.8% | Anthropic, Claude Opus 4 and Claude Sonnet 4checked 2026-10-15; printed on 1 more page | - |
Healthcare
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| HealthBench | 59.8 | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| HealthBench Hard | 31.6 | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| HealthBench Hard (hallucination rate)lower is better | 12.9% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Safety and Alignment
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OpenAI deception evals (deception rate)lower is better | 22% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked.
| Price | Value | Printed by | Setup as printed |
|---|---|---|---|
| Input | $2 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input | $0.5 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output | $8 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Input, batch | $1 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output, batch | $4 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Input, flex | $1 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input, flex | $0.25 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output, flex | $4 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Input, fast | $3.5 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input, fast | $0.875 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output, fast | $14 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
More from OpenAI
- GPT-6.1 Terra
- GPT-6.1 Sol
- GPT-6 Luna
- GPT-6 Sol
- GPT-6 Astra
- GPT-5.6 Luna
- GPT-5.6 Sol
- GPT-5.6 Terra
- GPT-5.5
- GPT-5.5 Pro
- GPT-5.4 mini
- GPT-5.4 nano
- GPT-5.4
- GPT-5.4 Pro
- GPT-5.3-Codex
- GPT-5.2-Codex
- GPT-5.2
- GPT-5.2 Instant
- GPT-5.2 Pro
- GPT-5.1-Codex-Max
- GPT-5.1
- GPT-5.1 Instant
- GPT-5-Codex
- GPT-5
- GPT-5 mini
- GPT-5 Pro
- gpt-oss-120b
- gpt-oss-20b
- ChatGPT agent
- codex-1
- OpenAI o4-mini
- GPT-4.1
- GPT-4.5
- Deep research
- o3-mini
- o1
- o1-mini
- GPT-4o
- GPT-5.5 Instant
- gpt-5.6-cyber
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.