Claude Sonnet 4.6 vs GPT-5.4
Benchmark Performance
Available Benchmarks
| Benchmark | Claude Sonnet 4.6 | GPT-5.4 |
|---|---|---|
| ARC-AGI-1verified-v1-15fb467fd4fc · verified_score · leader | 86.5092% of row best · percent · Claude Sonnet 4.6 (High) | 93.67100% of row best · percent · GPT-5.4 (XHigh) |
| ARC-AGI-2verified-v2-9a3db289984e · verified_score · leader | 60.4282% of row best · percent · Claude Sonnet 4.6 (High) | 73.95100% of row best · percent · GPT-5.4 (XHigh) |
| LMArena Agent Arenaagent-2026-08-31-011508720696 · outcome_score · leader | 0.9598% of row best · score · Claude Sonnet 4.6; 95% CI [-0.31540788, 2.21833799]; sessions 38385; observations 1371137; rank 31 | 3.25100% of row best · score · GPT 5.4 (High); 95% CI [2.38938480, 4.10626048]; sessions 76973; observations 3360684; rank 21 |
| LMArena Document Arenadocument-2026-07-30-011508720696 · arena_rating · statistical tie | 1,482.80100% of row best · rating · claude-sonnet-4-6; 95% CI [1476.49799837, 1489.09368537]; votes 54427; rank 8 | 1,470.1499% of row best · rating · gpt-5.4; 95% CI [1463.54357191, 1476.73400825]; votes 29809; rank 15 |
| LMArena Search Arenasearch-2026-08-24-011508720696 · arena_rating · leader | 1,221.23100% of row best · rating · claude-sonnet-4-6-search; 95% CI [1216.39472071, 1226.06821163]; votes 134905; rank 7 | 1,197.1698% of row best · rating · gpt-5.4-search; 95% CI [1191.73811150, 1202.57682579]; votes 110116; rank 16 |
| LMArena Text Arenatext-2026-09-01-011508720696 · arena_rating · statistical tie | 1,458.01100% of row best · rating · claude-sonnet-4-6; 95% CI [1454.41608210, 1461.60383515]; votes 66343; rank 32 | 1,452.51100% of row best · rating · gpt-5.4; 95% CI [1448.73439083, 1456.28324387]; votes 63615; rank 37 |
| LMArena Vision Arenavision-2026-08-27-011508720696 · arena_rating · statistical tie | 1,282.2599% of row best · rating · claude-sonnet-4-6; 95% CI [1275.84615489, 1288.65223369]; votes 25610; rank 25 | 1,293.18100% of row best · rating · gpt-5.4; 95% CI [1286.41064849, 1299.95761489]; votes 21245; rank 20 |
| LiveBench2026-06-25 · overall · leader | 77.1194% of row best · percent · claude-sonnet-4-6-thinking-auto-medium-effort · 13,226 output tokens / case | 82.38100% of row best · percent · gpt-5.4-xhigh · 18,273 output tokens / case |
| ToneBench2026-08-28-10-task-cd9819ab6e4d · overall_score · leader | 83.63100% of row best · points · Claude Sonnet 4.6 · 2,525 output tokens / case | 82.4799% of row best · points · GPT-5.4 · 3,254 output tokens / case |
| Overall ResultCounted from the protocol-matched rows above · 3 ties | 2 benchmark wins | 4 benchmark winsOverall lead |
Third-party benchmark Only like-for-like primary-publisher results are shown; raw scores, relative scores, configuration, and token spend remain visible.
Technical Differences
Side-by-Side Facts
| Field | Claude Sonnet 4.6 | GPT-5.4 |
|---|---|---|
| Developer | Anthropic | Openai |
| Family | Claude 4 | Gpt 5 4 |
| Model | Claude Sonnet 4.6 | GPT-5.4 |
| Version | Claude Sonnet 4.6 | GPT-5.4 |
| Lifecycle | active | active |
| Released | 2026-02-17 | 2026-03-05 |
| Knowledge cutoff | 2025-08-01 | 2025-08-31 |
| Input modalities | Unknown | Unknown |
| Output modalities | Unknown | Unknown |
| Context window | 1,000,000 | 1,050,000 |
| Total parameters | Unknown | Unknown |
| Active parameters | Unknown | Unknown |
| License | Unknown | Unknown |
| Open weights | No | No |
| API available | Yes | Yes |
| Self-hostable | No | No |
| Provider access | Anthropic (Standard), Deepinfra (Standard) | Openai (Standard), Openrouter (Standard) |
| Capabilities | chat, generation, reasoning, structured_outputs, tools | chat, generation, reasoning, structured_outputs, tools |
13 comparable fields · 8 material differences · Interactive comparison only; indexing gate not met
Claude Sonnet 4.6 Capabilities
GPT-5.4 Capabilities
Internal Comparison Graph
Related Comparisons
| A | Pair | B | Context |
|---|---|---|---|
GPT-5.4OpenAI | vs | GPT-5.5OpenAI | family variantsimage, text |
GPT-5.3-CodexOpenAI | vs | GPT-5.4OpenAI | family variantsimage, text |
Claude Sonnet 4.5Anthropic | vs | Claude Sonnet 4.6Anthropic | family variantsimage, text |
Claude Opus 4.6Anthropic | vs | Claude Sonnet 4.6Anthropic | family variantsimage, text |
Claude Opus 4.7Anthropic | vs | Claude Sonnet 4.6Anthropic | family variantsimage, text |
Claude Sonnet 4.6Anthropic | vs | Claude Sonnet 5Anthropic | family variantsimage, text |
Claude Sonnet 4.6Anthropic | vs |  Qwen3.7 FlashQwen | cross-developer peersimage, text |
Claude Sonnet 4.6Anthropic | vs | Grok 4.3xAI | cross-developer peersimage, text |
Claude Opus 4.6Anthropic | vs | GPT-5.4OpenAI | cross-developer peersimage, text |
Claude Opus 4.7Anthropic | vs | GPT-5.4OpenAI | cross-developer peersimage, text |
GPT-4.1OpenAI | vs | GPT-5.4OpenAI | family variantsimage, text |
GPT-4.1 MiniOpenAI | vs | GPT-5.4OpenAI | family variantsimage, text |
Primary Evidence
Sources and Freshness
Questions
Claude Sonnet 4.6 vs GPT-5.4 FAQs
Is Claude Sonnet 4.6 or GPT-5.4 better for coding?+
This comparison does not currently contain a protocol-matched coding benchmark for both Claude Sonnet 4.6 and GPT-5.4, so Model Markets cannot name a coding leader from pricing, context size, or capability labels alone.
Which is cheaper, Claude Sonnet 4.6 or GPT-5.4?+
Claude Sonnet 4.6 is $3.00 and GPT-5.4 is $2.50 per million tokens, so GPT-5.4 is cheaper on this metric. Claude Sonnet 4.6 is $15.00 and GPT-5.4 is $15.00 per million tokens, so they are tied on this metric.
Which has a larger context window, Claude Sonnet 4.6 or GPT-5.4?+
GPT-5.4 has the larger sourced context window. Claude Sonnet 4.6 supports 1,000,000 and GPT-5.4 supports 1,050,000.
Which performs better in benchmarks, Claude Sonnet 4.6 or GPT-5.4?+
GPT-5.4 leads the current overall benchmark count. The result uses 9 protocol-matched benchmarks from 4 publishers; it is not a universal quality score.
Can Claude Sonnet 4.6 or GPT-5.4 be self-hosted?+
Both models have the same recorded self-hosting status: unsupported. Claude Sonnet 4.6 is not marked open weight; GPT-5.4 is not marked open weight.
Can Claude Sonnet 4.6 and GPT-5.4 understand images?+
Claude Sonnet 4.6 is not documented with image input; GPT-5.4 is not documented with image input. This reflects supported input modalities, not vision quality.
Which can generate longer answers, Claude Sonnet 4.6 or GPT-5.4?+
Neither has a larger sourced maximum output. Claude Sonnet 4.6 is 128,000 and GPT-5.4 is 128,000.
Do Claude Sonnet 4.6 and GPT-5.4 support reasoning and tool use?+
Claude Sonnet 4.6: reasoning and tool calling. GPT-5.4: reasoning and tool calling. Feature support does not establish relative quality.
Which is available from more inference providers, Claude Sonnet 4.6 or GPT-5.4?+
Claude Sonnet 4.6 has 2 sourced provider routes; GPT-5.4 has 2, a tie.
Which offers better value, Claude Sonnet 4.6 or GPT-5.4?+
There is no universal value winner. Compare the input and output prices above with the matched benchmark result for your workload: cheaper tokens can be offset by different quality, token usage, latency, or provider availability.