Claude benchmark scores live in three places that disagree: Anthropic release notes, Scale's SEAL leaderboard, and third-party trackers. The disagreement is scaffolding, not error. This page consolidates every current score with the exact API model ID and price per row, and labels which harness produced each number.
The current lineup as of September 1, 2026: Fable 5 at $10/$50 per MTok, Opus 5 at $5/$25 (GA July 24), Sonnet 5 at $2/$10 (GA June 30), and Haiku 4.5 at $1/$5. Fable 5 leads SWE-bench Verified at 95.0% and SWE-bench Pro at 80.0%. Opus 5 has no published SWE-bench number: Anthropic reported Frontier-Bench v0.1 and CursorBench 3.2 at launch, and the only standardized third-party score for it is 89.1% on Terminal-Bench 2.1, 0.4 points behind GPT-5.6 Sol. Sonnet 5 scores 85.2% Verified and 63.2% Pro. Opus 4.8 (88.6% Verified, 69.2% Pro, 74.6% Terminal-Bench 2.1) is still active and retires no sooner than May 28, 2027, so the previous generation remains the reference point for anyone who needs a SWE-bench figure at the Opus price. Every table below carries the model ID, the price, and the harness that produced the number.
Master Table: Claude Models Comparison, Every Current Claude Model
Anthropic's current lineup is four models: Fable 5 (GA June 9, 2026), Opus 5 (GA July 24, 2026), Sonnet 5 (GA June 30, 2026), and Haiku 4.5. The 4.x generation below them stays active on the API rather than retiring with the launch, so Opus 4.8, 4.7, 4.6, 4.5, Sonnet 4.6, and Sonnet 4.5 all still take requests. Mythos 5 is the same underlying model as Fable 5 with safety classifiers lifted, restricted to approved Project Glasswing partners. Fable 5 has been available again since July 1, 2026; an August 7, 2026 change to its biology classifiers cut fallbacks about 85%, and the queries it still declines are re-served by Opus 5.
| Model | API Model ID | SWE-bench Verified | SWE-bench Pro | Context / Max Output | $/MTok In / Out |
|---|---|---|---|---|---|
| Claude Fable 5 | claude-fable-5 | 95.0% | 80.0% | 1M / 128k | $10 / $50 |
| Claude Opus 5 | claude-opus-5 | not published | not published | 1M / 128k | $5 / $25 |
| Claude Sonnet 5 | claude-sonnet-5 | 85.2% | 63.2% | 1M / 128k | $2 / $10 |
| Claude Mythos 5 | claude-mythos-5 (restricted) | 93.9%* | 77.8%* | 1M / 128k | $10 / $50 |
| Claude Opus 4.8 | claude-opus-4-8 | 88.6% | 69.2% | 1M / 128k | $5 / $25 |
| Claude Opus 4.7 | claude-opus-4-7 | 87.6% | 64.3% | 1M / 128k | $5 / $25 |
| Claude Opus 4.6 | claude-opus-4-6 | 80.8% | 51.9%† | 1M / 128k | $5 / $25 |
| Claude Sonnet 4.6 | claude-sonnet-4-6 | 79.6% | n/a | 1M / 64k | $3 / $15 |
| Claude Haiku 4.5 | claude-haiku-4-5 | 73.3% | 39.5%† | 200k / 64k | $1 / $5 |
| Claude Opus 4.5 (legacy) | claude-opus-4-5-20251101 | 80.9% | 45.9%† | 200k / 64k | $5 / $25 |
| Claude Sonnet 4.5 (legacy) | claude-sonnet-4-5-20250929 | 77.2% | 43.6%† | 200k / 64k | $3 / $15 |
SWE-bench Pro scores without a dagger are vendor-run. † = Scale SEAL leaderboard (standardized scaffolding), which runs well below vendor scaffolds. * = Claude Mythos Preview scores; Mythos 5 itself is not on public leaderboards. Opus 5 has no SWE-bench entry on either tracker; its numbers are in the Opus 5 section below. Verified and Pro scores read from the llm-stats trackers on September 1, 2026, plus provider announcements.
SWE-bench Verified Scores
SWE-bench Verified contains 500 human-validated Python tasks from real GitHub repositories. Models are scored on the percentage of tasks resolved correctly. Frontier scores are self-reported with provider scaffolds, so treat differences under ~1 point as noise.
| Rank | Model | Score | Provider |
|---|---|---|---|
| 1 | Claude Fable 5 | 95.0% | Anthropic |
| 2 | Claude Mythos Preview | 93.9% | Anthropic |
| 3 | Claude Opus 4.8 | 88.6% | Anthropic |
| 4 | Claude Opus 4.7 | 87.6% | Anthropic |
| 5 | Claude Sonnet 5 | 85.2% | Anthropic |
| 6 | Claude Opus 4.5 | 80.9% | Anthropic |
| 7 | Claude Opus 4.6 | 80.8% | Anthropic |
| 8 | DeepSeek-V4-Pro-Max | 80.6% | DeepSeek |
| 8 | Gemini 3.1 Pro | 80.6% | |
| 10 | MiniMax M3 | 80.5% | MiniMax |
| 11 | Qwen3.7 Max | 80.4% | Alibaba |
| 12 | Kimi K2.6 | 80.2% | Moonshot AI |
Source: llm-stats SWE-bench Verified tracker, read September 1, 2026 (111 models listed). Claude Opus 5 has no entry.
Seven of the top eight entries are Claude, and Sonnet 5 at 85.2% now sits above both Opus 4.5 and Opus 4.6 while costing $2/$10. The compression below that is the other half of the picture: DeepSeek V4 Pro Max, Gemini 3.1 Pro, MiniMax M3, Qwen3.7 Max, and Kimi K2.6 sit within 0.5 points of each other at ~80.4%, roughly where Opus 4.5 and 4.6 were a generation ago. Verified has a known contamination history and the 80% band is saturated. The 85% to 95% range above it is where the benchmark still differentiates, and Opus 5 is absent from it because Anthropic did not report a SWE-bench score for that model at all.
SWE-bench Pro Scores: Vendor vs Standardized
SWE-bench Pro contains 1,865 tasks across 41 professional repositories, split into public, commercial (private), and held-out sets. Two score families exist and they are not comparable: Anthropic's vendor-run numbers (own scaffold) and Scale's SEAL leaderboard (standardized scaffold, same harness for every model). Vendor scaffolds run 15 to 30 points higher.
Anthropic-Run Scores (Vendor Scaffold)
| Model | Score |
|---|---|
| Claude Fable 5 | 80.0% |
| Claude Mythos Preview | 77.8% |
| Claude Opus 4.8 | 69.2% |
| Qwen3.8 Max | 67.7% |
| Grok 4.5 | 64.7% |
| GPT-5.6 Sol | 64.6% |
| Claude Opus 4.7 | 64.3% |
| Claude Sonnet 5 | 63.2% |
| GPT-5.5 | 58.6% |
| Gemini 3.1 Pro | 54.2% |
Source: llm-stats SWE-bench Pro tracker, read September 1, 2026 (55 models, mean score 57.0%). Anthropic's own launch figure for Fable 5 was 80.3%.
Scale SEAL Leaderboard (Standardized Scaffold, Public Set)
Scale runs every model through the same harness, which isolates model capability from scaffold engineering. The newest Claude models (4.7, 4.8, Fable 5) are not yet listed; Opus 4.6 is the top Claude entry.
| Rank | Model | Score | CI |
|---|---|---|---|
| 1 | GPT-5.4 (xHigh) | 59.1% | ±3.56 |
| 2 | Muse Spark | 55.0% | ±3.60 |
| 3 | Claude Opus 4.6 (thinking) | 51.9% | ±3.61 |
| 4 | Gemini 3.1 Pro (thinking) | 46.1% | ±3.60 |
| 5 | Claude Opus 4.5 | 45.9% | ±3.60 |
| 6 | Claude Sonnet 4.5 | 43.6% | ±3.60 |
| 7 | Gemini 3 Pro | 43.3% | ±3.60 |
| 8 | Claude Sonnet 4 | 42.7% | ±3.59 |
| 9 | GPT-5 (High) | 41.8% | ±3.49 |
| 10 | GPT-5.2 Codex | 41.0% | ±3.57 |
| 11 | Claude Haiku 4.5 | 39.5% | ±3.55 |
| 12 | Qwen3 Coder 480B | 38.7% | ±3.55 |
On the harder private (commercial) set, the ordering flips: Claude Opus 4.6 (thinking) leads at 47.1%, ahead of Muse Spark (44.7%) and GPT-5.4 xHigh (43.4%), with Gemini 3.1 Pro at 32.2%. Claude degrades less than competitors when moving from public to unseen commercial repositories, which is the closer proxy for production codebases.
Sources: SEAL public leaderboard and private leaderboard, June 9, 2026.
The same Opus 4.6 scores 51.9% on Scale's standardized harness and materially higher on vendor scaffolds. Search and context retrieval are the usual difference. In Morph internal benchmarks, adding WarpGrep as a search subagent lifted Opus 4.6 from 55.4% to 57.5% on SWE-bench Pro while cutting cost 15.6% and latency 28%. WarpGrep runs in its own context window and issues up to 8 parallel tool calls per turn. Pricing: $0.80 per 100K tokens.
Opus 5 Benchmarks: Terminal-Bench 2.1, Frontier-Bench, and CursorBench
Claude Opus 5 shipped July 24, 2026 at $5/$25 per MTok, the same price as Opus 4.8, and it is the default model on Claude Max and the strongest model on Claude Pro. Anthropic did not report a SWE-bench score for it. The launch used Frontier-Bench v0.1, CursorBench 3.2, ARC-AGI 3, Zapier AutomationBench, OSWorld 2.0, GDPval-AA v2, HLE, and DeepSearchQA, and most of those were published as cost-performance curves across effort levels rather than single numbers.
| Benchmark | Opus 5 result | Who ran it |
|---|---|---|
| Terminal-Bench 2.1 | 89.1% (adaptive reasoning, max effort) | Artificial Analysis, Terminus 2 harness |
| Frontier-Bench v0.1 | State of the art, more than 2x Opus 4.8 at a lower cost per task | Anthropic |
| CursorBench 3.2 | Within 0.5% of Fable 5's peak at max effort, half the cost per task | Anthropic |
| ARC-AGI 3 | 3x the next-best model | Anthropic |
| Zapier AutomationBench | ~1.5x the next-best pass rate at the same cost per task | Anthropic |
| OSWorld 2.0 | Beats Fable 5's best result at just over a third of the cost | Anthropic |
| SWE-bench Verified / Pro | Not published | n/a |
Terminal-Bench 2.1 is the one standardized third-party number for Opus 5: 89.1% at max effort, second behind GPT-5.6 Sol at xhigh (89.5%) and ahead of Grok 4.6 at high (88.4%). Artificial Analysis runs the 89-task set in the Terminus 2 harness in an e2b sandbox and reports pass@1 averaged over three repeats. That is a 14.5-point jump over Opus 4.8's 74.6% on the same benchmark version, on the axis where Claude used to trail OpenAI.
Two specifics matter for routing. Opus 5's reliable knowledge cutoff is May 2026, four months later than Fable 5 and the 4.x models at January 2026, which shows up on questions about recent libraries and APIs. And Fable 5's safety classifiers fall back to Opus 5 rather than to a 4.x model, so Opus 5 is what answers a blocked Fable request. Anthropic states Opus 5 remains behind Mythos 5 on cybersecurity tasks. Specs: 1M context, 128k max output, adaptive thinking with default effort high, retirement no sooner than July 24, 2027.
Sources: Anthropic, Introducing Claude Opus 5 (July 24, 2026), Artificial Analysis Terminal-Bench v2.1, and the Claude models overview, all read September 1, 2026.
Sonnet 5 Benchmarks and Pricing
Claude Sonnet 5 went GA June 30, 2026 and is the default model on the Free and Pro plans. It scores 85.2% on SWE-bench Verified and 63.2% on SWE-bench Pro, which puts it above Opus 4.6 (80.8% Verified) and within 1.1 points of Opus 4.7 on Pro at a fifth of the input price. The full spec sheet, rate limits, and API examples live on the Claude Sonnet 5 page; what follows is how it places inside the Claude lineup.
| Sonnet 5 | Sonnet 4.6 | Opus 4.8 | |
|---|---|---|---|
| API model ID | claude-sonnet-5 | claude-sonnet-4-6 | claude-opus-4-8 |
| SWE-bench Verified | 85.2% | 79.6% | 88.6% |
| SWE-bench Pro | 63.2% | n/a | 69.2% |
| Context / max output | 1M / 128k | 1M / 64k | 1M / 128k |
| $/MTok in / out | $2 / $10 | $3 / $15 | $5 / $25 |
| Reliable knowledge cutoff | Jan 2026 | Aug 2025 | Jan 2026 |
The price needs one correction that Anthropic's own launch charts still carry: Sonnet 5 launched with introductory pricing of $2/$10 against a standard $3/$15, and on August 10, 2026 Anthropic made the introductory rate permanent. Any comparison built on the $3/$15 figure overstates its cost by 50%. Sonnet 4.6 stays active at $3/$15, so the cheaper model is also the stronger one within the Sonnet line.
Sources: Anthropic, Introducing Claude Sonnet 5 (June 30, 2026, price note edited August 10, 2026) and the llm-stats trackers, read September 1, 2026.
Opus 4.8 vs Opus 4.7: What Changed
Opus 4.8 (released May 28, 2026) is a same-price upgrade over 4.7. Anthropic's release notes claim it is 4x less likely to let flaws pass in code review.
| Benchmark | Opus 4.8 | Opus 4.7 |
|---|---|---|
| SWE-bench Pro | 69.2% | 64.3% |
| SWE-bench Verified | 88.6% | 87.6% |
| GPQA Diamond | 93.6% | n/a |
| OSWorld | 83.4% | 82.3% (OSWorld-Verified) |
| GDPval-AA (Elo) | 1890 | n/a |
| Price ($/MTok in / out) | $5 / $25 | $5 / $25 |
| Fast mode ($/MTok in / out) | $10 / $50 | $30 / $150 |
| Training cutoff | Jan 2026 | Jan 2026 |
The fast-mode repricing is the underreported change: Opus 4.8 fast mode (research preview, ~2.5x faster output) costs $10/$50 per MTok, a third of the $30/$150 fast-mode price on Opus 4.6 and 4.7. On GDPval-AA, Opus 4.8's 1890 Elo compares to 1769 for GPT-5.5. One tokenizer caveat applies to both 4.7 and 4.8: models from Opus 4.7 onward (including Fable 5) use a new tokenizer that can produce up to 35% more tokens for the same text than pre-4.7 models, so per-request cost comparisons against 4.6 need re-baselining, not just a price-table read.
Terminal-Bench Scores
Terminal-Bench tests whether a model can operate a live terminal: environment management, debugging, multi-step system operations. It is the benchmark where Claude has historically trailed OpenAI, and the gap is narrowing rather than closed.
| Model | Version | Score |
|---|---|---|
| GPT-5.6 Sol (xhigh) | Terminal-Bench 2.1 | 89.5% |
| Claude Opus 5 (max effort) | Terminal-Bench 2.1 | 89.1% |
| Grok 4.6 (high) | Terminal-Bench 2.1 | 88.4% |
| GPT-5.5 | Terminal-Bench 2.1 | 78.2% |
| Claude Opus 4.8 | Terminal-Bench 2.1 | 74.6% |
| Claude Opus 4.6 | Terminal-Bench 2.0 | 65.4% |
For the grounding query that lands here: Claude Opus 4.6 scored 65.4% on Terminal-Bench 2.0, Opus 4.8 scores 74.6% on Terminal-Bench 2.1, and Opus 5 scores 89.1% on the same 2.1 task set at max effort. That last figure puts Claude 0.4 points behind GPT-5.6 Sol at xhigh (89.5%) after a generation of trailing by 3.6. Versions 2.0 and 2.1 use different task sets, so the 65.4% figure is not directly comparable to the rest. Scores come from Artificial Analysis running the 89-task set in the Terminus 2 harness, pass@1 averaged over three repeats.
GPQA, OSWorld, and Agentic Benchmarks
Coding benchmarks measure patch quality. These measure the reasoning and computer-operation capability underneath it.
| Benchmark | Score | What It Measures |
|---|---|---|
| GPQA Diamond | 93.6% | Graduate-level science questions validated by domain experts |
| OSWorld | 83.4% | Operating a real desktop OS through screenshots and actions |
| Online-Mind2Web | 84% | Live web-browsing task completion |
| GDPval-AA | 1890 Elo | Economically valuable knowledge work (GPT-5.5: 1769) |
Anthropic's announcement footnotes add that Opus 4.8 is the first model to break 10% on a legal agent benchmark scored on an all-tasks-pass standard. Fable 5 extends the frontier further on Anthropic's hardest internal evals: 29.3% on FrontierCode Diamond versus 13.4% for Opus 4.8, and 29.8% on GDP.pdf vision tasks versus 24.9% for GPT-5.5. Mythos 5, the classifier-lifted variant, posts 78.0% capture on ExploitBench and 46.1% on BioMysteryBench, which is why it is gated to vetted Project Glasswing partners rather than self-serve.
Pricing per Million Tokens (September 2026)
Benchmark tables without prices hide the real decision. Full Anthropic API price list, with cache and batch rates:
| Model | Input | Output | Cache Hit | Batch In / Out |
|---|---|---|---|---|
| Claude Fable 5 / Mythos 5 | $10.00 | $50.00 | $1.00 | $5.00 / $25.00 |
| Claude Opus 5 | $5.00 | $25.00 | $0.50 | $2.50 / $12.50 |
| Claude Opus 4.8 / 4.7 / 4.6 / 4.5 | $5.00 | $25.00 | $0.50 | $2.50 / $12.50 |
| Claude Sonnet 5 | $2.00 | $10.00 | $0.20 | $1.00 / $5.00 |
| Claude Sonnet 4.6 | $3.00 | $15.00 | $0.30 | $1.50 / $7.50 |
| Claude Haiku 4.5 | $1.00 | $5.00 | $0.10 | $0.50 / $2.50 |
| Claude Opus 4.1 / 4 (retired) | $15.00 | $75.00 | n/a | n/a |
Modifiers that apply across models: 5-minute cache writes bill at 1.25x base input and 1-hour writes at 2x; cache reads at 0.1x. The Batch API is 50% off both input and output. Setting inference_geo: "us" (Opus 4.6+ and Sonnet 4.6+) adds a 1.1x multiplier on all token categories. The web search tool costs $10 per 1,000 searches. There is no long-context surcharge on the 1M-window models: Fable 5, Mythos 5, Opus 5, Sonnet 5, Opus 4.8/4.7/4.6, and Sonnet 4.6 bill a 900k-token request at the same rate as a 9k one. Sonnet 5's $2/$10 started as introductory pricing and became permanent on August 10, 2026, which makes it cheaper than the older Sonnet 4.6 it outscores.
For context against competitors: GPT-5.5 is $5/$30 per MTok, gpt-5.4 is $2.50/$15, gpt-5.3-codex is $1.75/$14, and Gemini 3.1 Pro is $2/$12 for prompts up to 200k tokens ($4/$18 above). Opus 4.8 at $5/$25 sits between GPT-5.5 and gpt-5.4 on output price while leading both on SWE-bench Pro (69.2% vs 58.6% for GPT-5.5, Anthropic-run). See Anthropic API pricing for the full breakdown.
Claude API Model IDs and Versions (2026)
Exact strings to pass as model. From the 4.6 generation onward, IDs are dateless pinned snapshots; earlier models keep date suffixes.
| Model | API ID | Context | Max Output | Training Cutoff |
|---|---|---|---|---|
| Claude Fable 5 | claude-fable-5 | 1M | 128k | n/a (GA Jun 9, 2026) |
| Claude Opus 5 | claude-opus-5 | 1M | 128k | May 2026 |
| Claude Sonnet 5 | claude-sonnet-5 | 1M | 128k | Jan 2026 |
| Claude Mythos 5 | claude-mythos-5 | 1M | 128k | Glasswing partners only |
| Claude Opus 4.8 | claude-opus-4-8 | 1M (200k on MS Foundry) | 128k | Jan 2026 |
| Claude Opus 4.7 | claude-opus-4-7 | 1M | 128k | Jan 2026 |
| Claude Opus 4.6 | claude-opus-4-6 | 1M | 128k | Aug 2025 |
| Claude Sonnet 4.6 | claude-sonnet-4-6 | 1M | 64k | Jan 2026 (reliable: Aug 2025) |
| Claude Haiku 4.5 | claude-haiku-4-5-20251001 (alias: claude-haiku-4-5) | 200k | 64k | Jul 2025 |
| Claude Sonnet 4.5 | claude-sonnet-4-5-20250929 | 200k | 64k | n/a |
| Claude Opus 4.5 | claude-opus-4-5-20251101 | 200k | 64k | n/a |
The Batch API supports 300k output tokens on Opus 4.6+ and Sonnet 4.6 via the beta header output-300k-2026-03-24. Fable 5 is available on the Claude API, Claude Platform on AWS, Bedrock (anthropic.claude-fable-5), Vertex AI, and Microsoft Foundry.
Deprecation and Retirement Dates
| Model | API ID | Retires |
|---|---|---|
| Claude Sonnet 4 | claude-sonnet-4-20250514 | Retired June 15, 2026 |
| Claude Opus 4 | claude-opus-4-20250514 | Retired June 15, 2026 |
| Claude Opus 4.1 | claude-opus-4-1-20250805 | Retired August 5, 2026 |
| Claude Sonnet 4.5 | claude-sonnet-4-5-20250929 | No sooner than September 29, 2026 |
| Claude Haiku 4.5 | claude-haiku-4-5 | No sooner than October 15, 2026 |
| Claude Mythos Preview | claude-mythos-preview | Deprecated, replaced by claude-mythos-5 |
| Claude Haiku 3.5 | claude-3-5-haiku-20241022 | Retired (except Bedrock/Vertex) |
Which Claude Model Is Best for Coding?
Opus 5 is the default since July 24, 2026: 89.1% Terminal-Bench 2.1 at $5/$25 per MTok, the same price Opus 4.8 charges, and the default model on Claude Max. The catch for anyone comparing on SWE-bench is that Anthropic published no SWE-bench score for it, so Opus 4.8 (88.6% Verified, 69.2% Pro) is still the model to quote when a procurement doc asks for that number. Route by task shape:
| Task | Model | Why |
|---|---|---|
| Default coding agent | Opus 5 | 89.1% Terminal-Bench 2.1 at $5/$25; May 2026 knowledge cutoff |
| Published SWE-bench at Opus price | Opus 4.8 | 88.6% Verified, 69.2% Pro, $5/$25; fast mode at $10/$50 |
| Hardest long-horizon work | Fable 5 | 80.0% Pro, 95.0% Verified; 2x output price ($50/MTok) |
| Everyday edits, drafts, chat | Sonnet 5 | $2/$10, 85.2% Verified, same 1M context window |
| Code review, tests, subagents | Haiku 4.5 | $1/$5; 39.5% SEAL Pro beats GPT-5.2 Codex's private-set 27.7% |
| Migrations needing huge context | Opus 5 or Sonnet 5 | 1M tokens, no long-context surcharge |
Fable 5's 10.8-point SWE-bench Pro lead over Opus 4.8 costs 2x per output token, and it pays off on tasks where a failed run wastes more than the token delta: large refactors, overnight autonomous runs, frontier-difficulty bugs. Opus 5 sits underneath it on Anthropic's CursorBench 3.2 curve at half the cost per task, which is the trade most teams should take. For multi-agent setups, pairing the main model with a cheap search subagent beats upgrading the main model: that is the WarpGrep result above. Cross-vendor comparison lives at best AI model for coding.
Legacy Models: Claude 3.5 Sonnet and Earlier
Searches for Claude 3.5 Sonnet benchmarks still land here, so the official numbers, for the record: Claude 3.5 Sonnet (June 2024) scored 33.4% on SWE-bench Verified, 59.4% on GPQA Diamond, and 92.0% on HumanEval per Anthropic's launch announcement. The upgraded Claude 3.5 Sonnet (October 2024) raised SWE-bench Verified to 49.0%. Both were retired October 28, 2025; the recommended replacement is claude-sonnet-4-6.
| Model | Release | SWE-bench Verified |
|---|---|---|
| Claude 3.5 Sonnet | Jun 2024 | 33.4% |
| Claude 3.5 Sonnet (upgraded) | Oct 2024 | 49.0% |
| Claude 3.7 Sonnet | Feb 2025 | 62.3% |
| Claude Sonnet 4.5 | Sep 2025 | 77.2% |
| Claude Opus 4.5 | Nov 2025 | 80.9% |
| Claude Opus 4.6 | Feb 2026 | 80.8% |
| Claude Opus 4.7 | Early 2026 | 87.6% |
| Claude Opus 4.8 | May 2026 | 88.6% |
| Claude Fable 5 | Jun 2026 | 95.0% |
| Claude Sonnet 5 | Jun 2026 | 85.2% |
| Claude Opus 5 | Jul 2026 | Not published |
33.4% to 95.0% in two years. The 4.6 generation plateaued near 81% while Anthropic shipped 1M context and agentic tool use; 4.7 and 4.8 broke the plateau, and Fable 5 added another 6.4 points on top. Sonnet 5 landed at 85.2%, above every Opus before 4.7. The progression stops at Fable 5 because Opus 5 has no SWE-bench entry, which is the first time since 2024 that Anthropic's flagship coding model shipped without one. The spread on SWE-bench Pro stays wide (80.0% to 39.5% across current Claude models alone), which is why Pro is the better differentiator at the frontier.
Frequently Asked Questions
What Claude models are available in 2026 (Opus, Sonnet, Haiku versions)?
Claude Fable 5 (claude-fable-5, GA June 9, 2026, available again since July 1), Claude Opus 5 (claude-opus-5, GA July 24, 2026), Claude Sonnet 5 (claude-sonnet-5, GA June 30, 2026), Claude Haiku 4.5 (claude-haiku-4-5), and Claude Mythos 5 (claude-mythos-5, Project Glasswing partners only). The 4.x generation stays on the API: Opus 4.8, 4.7, 4.6, 4.5, Sonnet 4.6, and Sonnet 4.5. Sonnet 4 and Opus 4 retired June 15, 2026; Opus 4.1 retired August 5, 2026.
What are Claude Opus 5's benchmark scores?
Anthropic published no SWE-bench score for Opus 5. The July 24, 2026 launch reported Frontier-Bench v0.1 (state of the art, more than double Opus 4.8 at a lower cost per task), CursorBench 3.2 (within 0.5% of Fable 5's peak at max effort, half the cost per task), ARC-AGI 3 at 3x the next-best model, Zapier AutomationBench, and OSWorld 2.0. The standardized third-party number is Terminal-Bench 2.1: 89.1% at max effort, second to GPT-5.6 Sol xhigh (89.5%). Price is $5/$25 per MTok, matching Opus 4.8.
What is Claude Sonnet 5's SWE-bench score?
85.2% on SWE-bench Verified and 63.2% on SWE-bench Pro, above Sonnet 4.6's 79.6% Verified. Sonnet 5 costs $2/$10 per MTok after Anthropic made the introductory rate permanent on August 10, 2026, with a 1M-token context window and 128k max output.
What is the API model ID for Claude Sonnet 4.6?
claude-sonnet-4-6. No date suffix: from the 4.6 generation onward, Anthropic IDs are dateless pinned snapshots. 1M context, 64k max output, $3/$15 per MTok.
What is Claude Opus 4.8's SWE-bench Pro score?
69.2% in Anthropic's vendor-run evaluation, vs 64.3% for Opus 4.7, 58.6% for GPT-5.5, and 54.2% for Gemini 3.1 Pro. On Scale's standardized SEAL leaderboard, the top Claude entry is Opus 4.6 (thinking) at 51.9% public / 47.1% private; 4.7, 4.8, and Fable 5 are not yet listed there.
What is Claude Opus 4.6's Terminal-Bench score?
65.4% on Terminal-Bench 2.0. On the newer Terminal-Bench 2.1, Opus 4.8 scores 74.6% vs GPT-5.5's 78.2%. The two versions use different task sets and are not directly comparable.
What are Claude 3.5 Sonnet's official benchmark scores?
June 2024 release: 33.4% SWE-bench Verified, 59.4% GPQA Diamond, 92.0% HumanEval. October 2024 upgrade: 49.0% SWE-bench Verified. Retired October 28, 2025; replacement is claude-sonnet-4-6.
Which Claude model is best for coding?
Opus 5 by default (89.1% Terminal-Bench 2.1, $5/$25), or Opus 4.8 at the same price when you need a published SWE-bench figure (88.6% Verified, 69.2% Pro). Fable 5 for the hardest long-horizon work if 2x output price is justified (95.0% Verified, 80.0% Pro, $10/$50). Sonnet 5 for everyday coding at $2/$10. Haiku 4.5 for high-volume review and subagent work at $1/$5. In multi-agent setups, a specialized search subagent like WarpGrep lifts whichever main model you pick.
What is Claude Fable 5 and how does it score?
Fable 5 is Anthropic's tier above Opus, GA June 9, 2026 and available again since July 1, 2026: 95.0% SWE-bench Verified, 80.0% SWE-bench Pro (80.3% in Anthropic's own run), 29.3% FrontierCode Diamond (vs 13.4% for Opus 4.8). $10/$50 per MTok, 1M context, 128k max output, adaptive thinking always on. Requests its biology classifiers decline are re-served by Opus 5.
Do Claude models charge extra for long context?
No. Fable 5, Mythos 5, Opus 4.8/4.7/4.6, and Sonnet 4.6 include the full 1M-token window at standard per-token pricing. GPT-5.5 also offers 1M, while Gemini 3.1 Pro raises rates above 200k tokens ($2/$12 to $4/$18). One cost caveat: the tokenizer on Opus 4.7+ can produce up to 35% more tokens for the same text than pre-4.7 models.
Why do Scale SEAL scores differ from Anthropic's scores?
Scaffolding. Scale runs every model through one standardized harness on SWE-bench Pro's public set; Anthropic runs its own agent scaffold. Opus-class models score in the high 60s on vendor scaffolds and low 50s on Scale's. Use SEAL numbers to compare models against each other, vendor numbers to track a single vendor's generation-over-generation progress.
Build with Claude + WarpGrep
WarpGrep lifted every model it was paired with on SWE-bench Pro, taking Opus 4.6 from 55.4% to 57.5% while cutting cost 15.6% and latency 28%. It runs in its own context window and issues 8 parallel tool calls per turn. $0.80 per 100K tokens.