BlogLLM Benchmark Scores as Published, Side by Side
GPT-5: Benchmark Scores as Published
Every result the catalog holds for GPT-5: 55 tests in 14 areas, from 9 pages. 28 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 2025-08-07Context window 400KMax output 128KInput Text, imageOutput TextWeights ClosedData version 2026103001, as of 2026-10-30
Knowledge Work
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Vals Finance Agent (version not stated) | 46.9% | Anthropic, Claude Sonnet 4.5checked 2026-10-15 | - |
| OpenAI economically important tasks (wins or ties against experts) | 47.1% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 52.8% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 6 more pages | Effort: none |
| 74.9% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high | |
| SWE-bench Multilingual | 55.3% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: high; Harness: in-house harness derived from SWE-agent; Tools: enabled; bash and edit tools |
| Terminal-Bench 2.0 | 35.2% | DeepSeek, DeepSeek-V3.2 model cardchecked 2026-10-15; printed on 1 more page | Effort: high |
| Terminal-Bench (version not stated) | 43.8% | Anthropic, Claude Haiku 4.5checked 2026-10-15; printed on 2 more pages | - |
| Terminal-Bench Hard | 30.5% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | Effort: high |
| Multi-SWE-bench | 39.3% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Harness: in-house harness derived from SWE-agent; Tools: enabled; bash and edit tools |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 77.8% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 5 more pages | Effort: none; Tools: no tools |
| 87.3% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: python | |
| 85.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: no tools | |
| AIME 2025 | 61.9% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 6 more pages | Effort: none; Tools: no tools |
| 94.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: no tools | |
| 71% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none; Tools: python | |
| 99.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: python | |
| SciCode | 42.9% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Tools: none |
| HMMT February 2025 | 93.3% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 3 more pages | Effort: high; Tools: no tools |
| 96.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: python | |
| IMO-AnswerBench | 76% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: high; Tools: none |
| HMMT November 2025 | 89.2% | Z.ai, GLM-4.7 model cardchecked 2026-10-15 | Effort: high |
| FrontierMath Tier 1-3 | 26.3% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 1 more page | Effort: high; Tools: python |
| 13.5% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: no tools |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp Long Context 128k | 90% | OpenAI, Introducing GPT-5.1 for developerschecked 2026-10-15 | Effort: high; Context: 128k |
Multimodal Understanding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| CharXiv Reasoning (no tools) | 81.1% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 57.8% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none | |
| MMMU-Pro | 78.4% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 62.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none | |
| MMMU (validation) | 84.2% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 3 more pages | Effort: high |
| 74.4% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none | |
| Video-MMMU | 84.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 61.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none | |
| ERQA | 65.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 42% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 6.3% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 3 more pages | Effort: none; Tools: no tools |
| 24.8% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high; Tools: no tools | |
| MMLU-Pro | 87.1% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: high; Tools: none |
| Humanity's Last Exam (with tools) | 35.2% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 2 more pages | Effort: high; Tools: python + search with blocklist |
| MMLU-Redux | 95.3% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Tools: none |
| FActScore (hallucination rate)lower is better | 1% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| LongFact-Concepts (hallucination rate)lower is better | 0.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| LongFact-Objects (hallucination rate)lower is better | 0.8% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
Code Generation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LiveCodeBench v6 | 87% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: high; Tools: none |
| Codeforces rating | 2537 | DeepSeek, DeepSeek-V3.2 model cardchecked 2026-10-15 | Effort: high |
| Aider polyglot | 26.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none |
| 88% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high | |
| OJBench (C++) | 56.2% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Tools: none |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Toolathlon | 29% | DeepSeek, DeepSeek-V3.2 model cardchecked 2026-10-15 | Effort: high |
| τ²-bench Telecom | 38.6% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 3 more pages | Effort: none |
| 96.7% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high | |
| τ²-bench Retail | 72.8% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 3 more pages | Effort: none |
| 81.1% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high | |
| τ²-bench Airline | 62.6% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 3 more pages | Effort: high |
| 55% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none | |
| τ²-bench (domain not stated) | 80.2% | DeepSeek, DeepSeek-V3.2 model cardchecked 2026-10-15; printed on 1 more page | Effort: high |
Web Search and Research
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| BrowseComp | 54.9% | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 2 more pages | Effort: high |
| BrowseComp-ZH | 63% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: high; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| Seal-0 | 51.4% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| FRAMES | 86% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15; printed on 1 more page | Effort: high; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| FinSearchComp T3 | 48.5% | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Tools: enabled; search, code interpreter, web browsing; Context: 256000 |
| Research-Eval (Reka) | 45.5 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
| X Browse | 24.2 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
| FRAMES (average cost per task)lower is better | $0.058 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
| Research-Eval (Reka, average cost per task)lower is better | $0.107 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
| X Browse (average cost per task)lower is better | $0.198 | xAI, Grok 4.1 Fast and Agent Tools APIchecked 2026-10-15 | Tools: agentic search (web search, X search, code execution); exact tools per row not stated |
Instruction Following
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MultiChallenge | 54.9% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none |
| 69.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high | |
| COLLIE | 99% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 70.5% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none |
Multilingual
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MMMLU | 89.4% | Anthropic, Claude Haiku 4.5checked 2026-10-15; printed on 1 more page | - |
Writing and Conversation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Longform Writing | 71.4 | Moonshot AI, Kimi K2 Thinking release pagechecked 2026-10-15 | Effort: high; Tools: none |
Healthcare
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| HealthBench | 54.3 | OpenAI, Introducing GPT-5checked 2026-10-15; printed on 1 more page | Effort: none |
| 67.2 | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high | |
| HealthBench Hard | 46.2 | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 25.5 | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none | |
| HealthBench Hard (hallucination rate)lower is better | 1.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 3.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none |
Safety and Alignment
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OpenAI deception evals (deception rate)lower is better | 4.8% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: high |
| 11.6% | OpenAI, Introducing GPT-5checked 2026-10-15 | Effort: none |
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 | $1.25 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input | $0.125 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output | $10 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Input, batch | $0.625 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input, batch | $0.0625 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output, batch | $5 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Input, flex | $0.625 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input, flex | $0.0625 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output, flex | $5 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Input, fast | $2.5 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Cached input, fast | $0.25 | OpenAI, OpenAI API pricingchecked 2026-10-15 | - |
| Output, fast | $20 | 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 mini
- GPT-5 Pro
- gpt-oss-120b
- gpt-oss-20b
- ChatGPT agent
- codex-1
- o3
- 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.