Anthropic says Claude Opus 5 matches the performance of its pricier Fable 5 model at half the cost
Anthropic has released Claude Opus 5, a new model designed to compete with Fable 5 while costing significantly less. The company claims Opus 5 achieves near-Fable 5 performance across coding and knowledge work, but at half the token price.
This launch comes as Anthropic faces pricing pressure from competitors, including GPT-5.6 Sol and Chinese rivals. Opus 5 is now the default model for Claude Max subscribers and the most capable option for Claude Pro users.
The context window remains at one million tokens. Pricing for standard usage stays at $5 per million input tokens and $25 per million output tokens, matching the previous Opus 4.8 version. A new Fast Mode option increases speed by 2.5 times but doubles the cost.
Price comparison
| Model | Input Tokens | Cache Writes (5 min.) | Cache Writes (1 hr) | Cache Hits & Updates | Output tokens |
|---|---|---|---|---|---|
| Claude Fable 5 | $10 / MTok | $12.50 / MTok | $20 / MTok | $1 / MTok | $50 / MTok |
| Claude Mythos 5 (limited availability) | $10 / MTok | $12.50 / MTok | $20 / MTok | $1 / MTok | $50 / MTok |
| Claude Opus 5 | $5 / MTok | $6.25 / MTok | $10 / MTok | $0.50 / MTok | $25 / MTok |
Token rates alone do not tell the full story. Efficiency matters. Opus 4.7 ended up costing 30 to 40 percent more per task than Opus 4.6, despite identical base rates. A similar pattern appeared recently with Claude Sonnet 5.
Higher effort can cost more and perform worse
Users can trade performance against token use through five effort settings: low, medium, high, xhigh, and max. Anthropic says Opus 5 offers better value than its predecessor at every level.
The company’s prompting guide recommends using the “low” and “medium” settings broadly. These deliver good results with a fraction of the token use and latency while beating the same settings on earlier Opus models. For coding and agentic tasks, Anthropic still recommends starting with “xhigh.”
Opus 5 scores slightly worse at the max effort setting than at the second-highest setting on two benchmarks. The drop appears on Frontier-Bench v0.1 and the Artificial Analysis Coding Agent Index.
Opus 5 beats Fable 5 at agentic coding
According to Anthropic’s own benchmarks, Opus 5 sets records across several evaluations. On Frontier-Bench v0.1, the model hits 43.3 percent on agentic terminal coding. This beats Fable 5 (33.7 percent), GPT-5.6 Sol (34.4 percent), and its predecessor Opus 4.8 (21.1 percent) by wide margins.
On the knowledge work benchmark GDPval-AA v2, Opus 5 leads with an Elo score of 1,861. This is ahead of Fable 5 (1,747) and GPT-5.6 Sol (1,736).
Opus 5 does not win everywhere. On agentic coding via DeepSWE v1.1, GPT-5.6 Sol leads with 72.7 percent, followed by Fable 5 (69.7 percent) and Opus 5 (68.8 percent). On health tasks and legal benchmarks, Fable 5 and Mythos 5 outperform Opus 5, respectively.
The ARC-AGI-3 result is likely the biggest surprise in Anthropic’s benchmarks. The benchmark measures novel problem-solving without memorized patterns. Opus 5 scores 30.2 percent. Opus 4.8 managed just 1.5 percent, while GPT-5.6 Sol hit 7.8 percent. That is nearly a 4x gap over the next-best model. There is no Fable 5 result for this benchmark, and it is unclear whether such a large lead on a test will show up in actual use.
Opus 5 falls behind Mythos 5 on cybersecurity tasks. Anthropic says it deliberately did not train Opus 5 on cyber tasks, as was also the case with its predecessor. The model comes close to Mythos 5 at finding vulnerabilities but performs much worse when asked to exploit them.
Anthropic also says Opus 5 has improved at generating visual outputs and analyzing visual content such as charts and diagrams.
Opus 5 builds its own tools when it needs them
Anthropic describes Opus 5 as much better at checking its own work and improving it through iteration. In a Frontier-Bench task, Opus 5 received a drawing of a machine part and had to create a 3D model in FreeCAD. The catch was that the model intentionally had no way to view the drawing directly. Opus 5 responded by writing its own computer vision pipeline to extract the geometry from raw pixels, then reconstructed the complete machine part. No other model solved this task after five attempts, Anthropic says.
Opus 5 also worked on a real bug in a popular open-source package manager. According to Anthropic, it found the root cause and fixed an edge case that the community patch had missed. A competing model fixed only the surface symptom before reporting the bug as resolved.
An engineer at a trading firm reportedly used Opus 5 to build a market data feed for a new exchange in one session. Previous models could not complete the task, even with detailed plans.
Opus 5’s cyber filters trigger less often than Fable 5’s
Opus 5’s safety setup allows source code vulnerability research but blocks binary-based vulnerability scanning, penetration testing, and exploit generation, according to Anthropic. The cyber classifiers trigger about 85 percent less often than on Fable 5. Fable 5’s frequent interventions drew heavy criticism. Blocked requests in Claude.ai, Claude Code, and Claude Cowork default to Opus 4.8 as a fallback, the same approach used with Fable 5.
Anthropic calls Opus 5 the most capable generally available model for scientific research. It shows gains over Opus 4.8 across all life sciences evaluations, with standout improvements in organic chemistry (plus 10.2 percentage points on deriving molecular structures from spectroscopy data) and protein-related tasks (plus 7.7 percentage points).
Alongside Opus 5, Anthropic is releasing two beta features. Mid-Conversation Tool Changes on the Claude Platform let developers swap available tools during a conversation without invalidating the prompt cache. Automatic Fallbacks on the API route blocked requests to a different model automatically.
