Claude Opus 4.8
Not tested yet
Language modelLegacy
Claude Opus 4.8
Anthropic
- Released
- May 2026
- Data date
- July 30, 2026
- Intelligence
- Overall score
- Price
- Input · Output / 1M tokens
Methodology and source
The index averages rank percentiles from 4 documented comparison cohorts.
Entries 1-12 of 15. Page 1 of 2.
Published benchmarks
Each chart compares models in the same documented test and highlights the current model.
Reasoning
2 measurementsGPQA Diamond
Overall
Note (Claude Opus 4.8): Editorial selection from the vendor table, not a complete extract of the comparison cohort.
Higher is better2 modelsClaude Opus 4.8: Rank 2 of 2Data date: September 2, 2026
1.Claude Opus 4.794.2%
Note: Editorial selection from the vendor table, not a complete extract of the comparison cohort.
Scale: 0% - 100%
Methodology and context
- Comparison cohort
- gpqa-anthropic-opus-4-8-release
- Metric
- accuracy
- Scale
- 0-100
- Participants
- 2
- Methodology and context
- Vendor-reported GPQA Diamond result.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: Anthropic. Dataset type: public. Leaderboard date: 2026-05-28. Source ID: anthropic. Evidence capture ID: curated-anthropic-2026-09-02-b68f06b7b001. Source payload hash: b68f06b7b001db1605515e7d19d67f6445e757877dfd9e4559e789fdb194744f. Parser version: curated-v1. Imported at: 2026-09-04
EnigmaEval
Overall
What this benchmark measures: 1,161 complex, often multimodal puzzles test creative problem solving beyond well-defined academic tasks.
Context: Puzzle solving is a narrow specialty and does not replace professional or everyday task evaluation.
Higher is better5 modelsClaude Opus 4.8: Rank 5 of 5Data date: August 26, 2026
Scale: 0% - 100%
Methodology and context
- Comparison cohort
- scale-enigma-eval-2026-07-23
- Metric
- pass@1 accuracy
- Scale
- 0-100
- Participants
- 5
- Methodology and context
- Complex multimodal puzzles with exact answer matching and one attempt.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Model snapshot: claude-opus-4.8-xhigh. Evaluator: Scale AI. Dataset type: public. Leaderboard date: 2026-07-23. Attempts: pass@1. Reasoning effort: xhigh. Score display: 23.51% ± 2.41. Source ID: scale-ai. Evidence capture ID: curated-scale-ai-2026-08-26-5ccace63d82b. Source payload hash: 5ccace63d82bb1236aeffd0c9d0760c126999320c83feb7ed8e4170382506715. Parser version: curated-v1. Imported at: 2026-09-04
Agent tasks
2 measurementsOSWorld Verified
Overall
What this benchmark measures: Computer tasks in realistic desktop environments.
Context: The result belongs to the complete agent system and is not an isolated model score.
Note (Claude Opus 4.8): Editorial selection from the vendor table, not a complete extract of the comparison cohort.
Higher is better2 modelsClaude Opus 4.8: Rank 1 of 2Data date: September 2, 2026
2.Claude Opus 4.782.3%
Note: Anthropic's release-page footnote reports 82.3% after an updated run. The associated system-card table still contains 82.8%.
Scale: 0% - 100%
Methodology and context
- Comparison cohort
- osworld-anthropic-opus-4-8-release
- Metric
- success
- Scale
- 0-100
- Participants
- 2
- Methodology and context
- Vendor-reported OSWorld Verified result for the agent system.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: Anthropic. Dataset type: public. Leaderboard date: 2026-05-28. Source ID: anthropic. Evidence capture ID: curated-anthropic-2026-09-02-7c06f0b8e422. Source payload hash: 7c06f0b8e422433920c25f65075e9d0e2acabcda53ec6285de0f64fa3603abfd. Parser version: curated-v1. Imported at: 2026-09-04
OSWorld 2.0
Overall
What this benchmark measures: 108 long computer workflows with up to 500 steps.
Context: It evaluates a complete system. Binary and partial success rates stay separate.
Note (Claude Opus 4.8): Claude used about 244,000 output tokens. This is a system result, not an isolated model score.
Higher is better2 modelsClaude Opus 4.8: Rank 1 of 2Data date: August 26, 2026
2.GPT-5.513%
Note: Claude used about 244,000 output tokens. This is a system result, not an isolated model score.
Scale: 0% - 100%
Methodology and context
- Comparison cohort
- osworld-2-v2026-06-24-long-horizon
- Metric
- binary accuracy
- Scale
- 0-100
- Participants
- 2
- Methodology and context
- Complete agent system with up to 500 steps.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: OSWorld 2.0. Dataset type: public. Leaderboard date: 2026-06-25. Tool policy: batched tool. Reasoning effort: max. Score display: 20.6% binary / 54.8% partial. Source ID: osworld. Evidence capture ID: curated-osworld-2026-08-26-a8d40127d291. Source payload hash: a8d40127d291171384cd4bf5ba7d84fff76fbac3ff5ddcfc570eb2f9f2bcf53d. Parser version: curated-v1. Imported at: 2026-09-04
BrowseComp
Overall
What this benchmark measures: 1,266 hard-to-find research tasks with short, unambiguous answers.
Context: The benchmark rewards persistent fact finding and only partly represents open-ended research reports or typical user questions.
Note (Claude Opus 4.8): Editorial selection from the vendor table, not a complete extract of the comparison cohort.
Higher is better6 modelsClaude Opus 4.8: Rank 5 of 6Data date: September 2, 2026
1.Claude Mythos 588%
Note: Editorial selection from the vendor table, not a complete extract of the comparison cohort.
2.GPT-5.6 Terra87.5%
Note: Editorial selection from the vendor table, not a complete extract of the comparison cohort.
3.Gemini 3.1 Pro Preview85.9%
Note: Editorial selection from the vendor table, not a complete extract of the comparison cohort.
4.GPT-5.584.4%
Note: Editorial selection from the vendor table, not a complete extract of the comparison cohort.
6.GPT-5.6 Luna83.3%
Note: Editorial selection from the vendor table, not a complete extract of the comparison cohort.
Scale: 0% - 100%
Methodology and context
- Comparison cohort
- browsecomp-openai-gpt-5-6-table
- Metric
- accuracy
- Scale
- 0-100
- Participants
- 6
- Methodology and context
- Cross-vendor comparison table with browsing tools. Exact agent systems differ.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: OpenAI. Dataset type: public. Leaderboard date: 2026-07-09. Source ID: openai. Evidence capture ID: curated-openai-2026-09-02-a3a0a1571322. Source payload hash: a3a0a15713220d01cc04e6ee194222775ffc1b80afe4ccb9b578912cebba533a. Parser version: curated-v1. Imported at: 2026-09-04
SWE-Marathon v1.0
Binary resolution, pass@1
What this benchmark measures: 20 multi-hour repository tasks test very long software work with binary end-to-end resolution.
Context: Scores belong to the v1.0 agent setup. No v1.1 leaderboard was posted at the data cutoff.
Higher is better6 modelsClaude Opus 4.8: Rank 2 of 6Data date: August 9, 2026
3.Claude Fable 524%
Note: Two refusals and three unavailable provider-ban trials were counted as zero.
Scale: 0% - 100%
Methodology and context
- Comparison cohort
- swe-marathon-v1-0:official:pass1
- Metric
- binary resolution pass@1
- Scale
- 0-100
- Participants
- 6
- Methodology and context
- 20 multi-hour tasks. The named coding agent is part of every result row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: SWE-Marathon. Dataset type: public. Leaderboard date: 2026-06-05. Agent: Claude Code. Source ID: swe-marathon. Evidence capture ID: curated-swe-marathon-2026-08-09-90e2ac9e0f51. Source payload hash: 90e2ac9e0f510d0877160a36eea31f579e7ba49a4b0d1117ec4cc50b37f5f787. Parser version: curated-v1. Imported at: 2026-09-04
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
SWE-Bench Pro: Overall (Show measurement, test conditions, and source)
- Source value
- 69.2
- Score
- 69.2
- Metric
- resolved
- Unit
- %
- Benchmark version
- Public
- Category
- coding
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- swe-pro-anthropic-opus-4-8-release
- Source type
- vendor-reported
- Evaluator
- Anthropic
- Status
- disputed
- Retrieved at
- 2026-09-02
- Methodology
- Comparison table published by Anthropic. Not a Gradually test.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: Anthropic. Dataset type: public-disputed. Leaderboard date: 2026-05-28. Source ID: anthropic. Evidence capture ID: curated-anthropic-2026-09-02-d0c8138a181c. Source payload hash: d0c8138a181c79d6502a11946b5b4b7752d056c9e6e3e363522f96a991029eff. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- Comparison table published by Anthropic. Not a Gradually test.
- Context
- After an audit, OpenAI estimates that about 30% of the public tasks are broken. The result therefore remains a disputed secondary signal. The rows are an editorial selection from the respective comparison table.. The source dataset marks this result as disputed. It remains visible as documented evidence, but is not reliable comparative proof.
TUA-Bench: Successful in all five runs (Show measurement, test conditions, and source)
- Source value
- 51.7
- Score
- 51.7
- Metric
- All-5
- Unit
- %
- Benchmark version
- paper and repository release
- Category
- agentic
- Task
- all5
- Task label
- Successful in all five runs
- Direction
- Higher is better
- Comparison cohort
- tua-bench:official:all5
- Source type
- official-leaderboard
- Evaluator
- TUA-Bench
- Status
- active
- Retrieved at
- 2026-08-09
- Methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: TUA-Bench. Dataset type: public. Leaderboard date: 2026-06-26. Agent: Claude Code. Reasoning effort: max. Source ID: tua-bench. Evidence capture ID: curated-tua-bench-2026-08-09-49b3d911f6b4. Source payload hash: 49b3d911f6b43226ff32707627d25d89b9f788f3fc0c3b32a5d81a659d2f84a2. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.
TUA-Bench: Successful in all five runs (Show measurement, test conditions, and source)
- Source value
- 45
- Score
- 45
- Metric
- All-5
- Unit
- %
- Benchmark version
- paper and repository release
- Category
- agentic
- Task
- all5
- Task label
- Successful in all five runs
- Direction
- Higher is better
- Comparison cohort
- tua-bench:official:all5
- Source type
- official-leaderboard
- Evaluator
- TUA-Bench
- Status
- active
- Retrieved at
- 2026-08-09
- Methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: TUA-Bench. Dataset type: public. Leaderboard date: 2026-06-26. Agent: OpenHands SDK. Reasoning effort: max. Source ID: tua-bench. Evidence capture ID: curated-tua-bench-2026-08-09-b67769f171df. Source payload hash: b67769f171df161ddf33bd6a20fff3a59f59b7c9cdcb89890fa3fc7e15731bf6. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.
TUA-Bench: Pass@1 (Show measurement, test conditions, and source)
- Source value
- 58.8
- Score
- 58.8
- Metric
- Pass@1
- Unit
- %
- Benchmark version
- paper and repository release
- Category
- agentic
- Task
- pass1
- Task label
- Pass@1
- Direction
- Higher is better
- Comparison cohort
- tua-bench:official:pass1
- Source type
- official-leaderboard
- Evaluator
- TUA-Bench
- Status
- active
- Retrieved at
- 2026-08-09
- Methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: TUA-Bench. Dataset type: public. Leaderboard date: 2026-06-26. Agent: Claude Code. Reasoning effort: max. Source ID: tua-bench. Evidence capture ID: curated-tua-bench-2026-08-09-cb3c6645cd6f. Source payload hash: cb3c6645cd6f574709685504d7d3df442d662cb6aa5bded5b9333feafa8be06a. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.
TUA-Bench: Pass@1 (Show measurement, test conditions, and source)
- Source value
- 57.3
- Score
- 57.3
- Metric
- Pass@1
- Unit
- %
- Benchmark version
- paper and repository release
- Category
- agentic
- Task
- pass1
- Task label
- Pass@1
- Direction
- Higher is better
- Comparison cohort
- tua-bench:official:pass1
- Source type
- official-leaderboard
- Evaluator
- TUA-Bench
- Status
- active
- Retrieved at
- 2026-08-09
- Methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: TUA-Bench. Dataset type: public. Leaderboard date: 2026-06-26. Agent: OpenHands SDK. Reasoning effort: max. Source ID: tua-bench. Evidence capture ID: curated-tua-bench-2026-08-09-ff1dfccddfa3. Source payload hash: ff1dfccddfa388a273a750a51d7189c0bbf70e9cf92281d408af61bc22b699c1. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.
TUA-Bench: Pass@5 (Show measurement, test conditions, and source)
- Source value
- 64.2
- Score
- 64.2
- Metric
- Pass@5
- Unit
- %
- Benchmark version
- paper and repository release
- Category
- agentic
- Task
- pass5
- Task label
- Pass@5
- Direction
- Higher is better
- Comparison cohort
- tua-bench:official:pass5
- Source type
- official-leaderboard
- Evaluator
- TUA-Bench
- Status
- active
- Retrieved at
- 2026-08-09
- Methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: TUA-Bench. Dataset type: public. Leaderboard date: 2026-06-26. Agent: Claude Code. Reasoning effort: max. Source ID: tua-bench. Evidence capture ID: curated-tua-bench-2026-08-09-c8f831dd7787. Source payload hash: c8f831dd778700a886b497a9324427ad6fb3088c4802a3ae2153b4fa4b6bfc1d. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- 120 tasks with five runs. Agent system and reasoning level are part of each row.
Recorded measurement details
Every observation retains its source value and published test conditions.
GPQA Diamond: Overall (Show measurement, test conditions, and source)
- Source value
- 93.6
- Score
- 93.6
- Metric
- accuracy
- Unit
- %
- Benchmark version
- Diamond
- Category
- reasoning
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- gpqa-anthropic-opus-4-8-release
- Source type
- vendor-reported
- Evaluator
- Anthropic
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- Vendor-reported GPQA Diamond result.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: Anthropic. Dataset type: public. Leaderboard date: 2026-05-28. Source ID: anthropic. Evidence capture ID: curated-anthropic-2026-09-02-b68f06b7b001. Source payload hash: b68f06b7b001db1605515e7d19d67f6445e757877dfd9e4559e789fdb194744f. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- Vendor-reported GPQA Diamond result.
- Context
- Editorial selection from the vendor table, not a complete extract of the comparison cohort.
OSWorld Verified: Overall (Show measurement, test conditions, and source)
- Source value
- 83.4
- Score
- 83.4
- Metric
- success
- Unit
- %
- Benchmark version
- Verified
- Category
- agentic
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- osworld-anthropic-opus-4-8-release
- Source type
- vendor-reported
- Evaluator
- Anthropic
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- Vendor-reported OSWorld Verified result for the agent system.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: Anthropic. Dataset type: public. Leaderboard date: 2026-05-28. Source ID: anthropic. Evidence capture ID: curated-anthropic-2026-09-02-7c06f0b8e422. Source payload hash: 7c06f0b8e422433920c25f65075e9d0e2acabcda53ec6285de0f64fa3603abfd. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- Vendor-reported OSWorld Verified result for the agent system.
- Context
- Editorial selection from the vendor table, not a complete extract of the comparison cohort.
BrowseComp: Overall (Show measurement, test conditions, and source)
- Source value
- 84.3
- Score
- 84.3
- Metric
- accuracy
- Unit
- %
- Benchmark version
- 1,266 tasks
- Category
- search
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- browsecomp-openai-gpt-5-6-table
- Source type
- vendor-reported
- Evaluator
- OpenAI
- Status
- active
- Retrieved at
- 2026-09-02
- Methodology
- Cross-vendor comparison table with browsing tools. Exact agent systems differ.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: OpenAI. Dataset type: public. Leaderboard date: 2026-07-09. Source ID: openai. Evidence capture ID: curated-openai-2026-09-02-a3a0a1571322. Source payload hash: a3a0a15713220d01cc04e6ee194222775ffc1b80afe4ccb9b578912cebba533a. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- Cross-vendor comparison table with browsing tools. Exact agent systems differ.
- Context
- Editorial selection from the vendor table, not a complete extract of the comparison cohort.
OSWorld 2.0: Overall (Show measurement, test conditions, and source)
- Source value
- 20.6
- Score
- 20.6
- Metric
- binary accuracy
- Unit
- %
- Benchmark version
- v2026.06.24
- Category
- agentic
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- osworld-2-v2026-06-24-long-horizon
- Source type
- official-leaderboard
- Evaluator
- OSWorld 2.0
- Status
- active
- Retrieved at
- 2026-08-26
- Methodology
- Complete agent system with up to 500 steps.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: OSWorld 2.0. Dataset type: public. Leaderboard date: 2026-06-25. Tool policy: batched tool. Reasoning effort: max. Score display: 20.6% binary / 54.8% partial. Source ID: osworld. Evidence capture ID: curated-osworld-2026-08-26-a8d40127d291. Source payload hash: a8d40127d291171384cd4bf5ba7d84fff76fbac3ff5ddcfc570eb2f9f2bcf53d. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- Complete agent system with up to 500 steps.
- Context
- Claude used about 244,000 output tokens. This is a system result, not an isolated model score.
EnigmaEval: Overall (Show measurement, test conditions, and source)
- Source value
- 23.51
- Score
- 23.51
- Metric
- pass@1 accuracy
- Unit
- %
- Benchmark version
- open-source release 2026-07-23
- Category
- reasoning
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Comparison cohort
- scale-enigma-eval-2026-07-23
- Confidence interval
- 2.41
- Source type
- official-leaderboard
- Evaluator
- Scale AI
- Status
- active
- Retrieved at
- 2026-08-26
- Methodology
- Complex multimodal puzzles with exact answer matching and one attempt.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Model snapshot: claude-opus-4.8-xhigh. Evaluator: Scale AI. Dataset type: public. Leaderboard date: 2026-07-23. Attempts: pass@1. Reasoning effort: xhigh. Score display: 23.51% ± 2.41. Source ID: scale-ai. Evidence capture ID: curated-scale-ai-2026-08-26-5ccace63d82b. Source payload hash: 5ccace63d82bb1236aeffd0c9d0760c126999320c83feb7ed8e4170382506715. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- Complex multimodal puzzles with exact answer matching and one attempt.
SWE-Marathon v1.0: Binary resolution, pass@1 (Show measurement, test conditions, and source)
- Source value
- 26
- Score
- 26
- Metric
- binary resolution pass@1
- Unit
- %
- Benchmark version
- v1.0 leaderboard
- Category
- coding
- Task
- pass1
- Task label
- Binary resolution, pass@1
- Direction
- Higher is better
- Comparison cohort
- swe-marathon-v1-0:official:pass1
- Source type
- official-leaderboard
- Evaluator
- SWE-Marathon
- Status
- active
- Retrieved at
- 2026-08-09
- Methodology
- 20 multi-hour tasks. The named coding agent is part of every result row.. Source model ID: curated/claude-opus-4.8. Source model name: Claude Opus 4.8. Provider reported by source: Anthropic. Evaluator: SWE-Marathon. Dataset type: public. Leaderboard date: 2026-06-05. Agent: Claude Code. Source ID: swe-marathon. Evidence capture ID: curated-swe-marathon-2026-08-09-90e2ac9e0f51. Source payload hash: 90e2ac9e0f510d0877160a36eea31f579e7ba49a4b0d1117ec4cc50b37f5f787. Parser version: curated-v1. Imported at: 2026-09-04
- Comparison methodology
- 20 multi-hour tasks. The named coding agent is part of every result row.
Our model tests
Community tests you can inspect
Compare how the models respond to the same prompt. Each test shows the first attempt, with no subsequent fixes to the generated code. These results do not contribute to an overall score.
Pelican on a bicycle
A community classic for free-form SVG drawing.
Prompt and test conditions
Generate an SVG of a pelican riding a bicycle
Token limit including reasoning: 8,192 tokens
Task origin (Simon Willison)Profile
Specifications and access
Published information about this model. Existing estimates are explicitly labeled.
Pricing
Published prices
Prices apply to the stated unit. Resolution, output length, and provider can change the cost.
Cost example
100 API requests
$1.75
100 requests with 1,000 input and 500 output tokens each calculate to $1.75. Input accounts for $0.5; output accounts for $1.25.
- Input price
- $5 / 1M tokensBase tier for each request
- Output price
- $25 / 1M tokensBase tier for each request
The calculation uses documented token prices and no cache discount. It excludes extra tools, tax, reasoning tokens, and further hidden output tokens.
Open sourceHead-to-head comparisons
Compare this model
Each matchup compares this model with exactly one other model from the same category.
Show all comparisons (42)
Evidence
Sources and data date
Every statement links to its underlying documentation or leaderboard.
- Price
Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs.
- Model metadata (retrieved September 2, 2026)Source
- Knowledge cutoff dataSource
- API pricingSource
- Anthropic (retrieved September 2, 2026)Source
- OpenAI (retrieved September 2, 2026)Source
- OSWorld 2.0 (retrieved August 26, 2026)Source
- Scale AI (retrieved August 26, 2026)Source
- SWE-Marathon (retrieved August 9, 2026)Source
- TUA-Bench (retrieved August 9, 2026)Source