Claude Opus 4.8 vs Step 3.7 Flash
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Math · Long documents
Wins 1 of 7 areas
Following instructions
Claude Opus 4.8 is the stronger all-rounder.Step 3.7 Flash is cheaper and better at following instructions.
Scores updated · 21 tests both models report · How we compare
Where each one wins
Tests won in each of the seven areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software40Claude Opus 4.84 of 4 tests
- AgentsCarrying out multi-step tasks on its own30Claude Opus 4.83 of 3 tests
- ReasoningHard problems that need careful thinking30Claude Opus 4.83 of 3 tests
- FactsGetting facts right instead of making them up20Claude Opus 4.82 of 2 tests
- MathCompetition and research-level math20Claude Opus 4.82 of 2 tests
- Long documentsFinding answers in very long texts10Claude Opus 4.81 of 1 test
- Following instructionsDoing exactly what it is asked01Step 3.7 Flash1 of 1 test
Long documents and following instructions rest on a single test each.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Step 3.7 Flash costs 96% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Opus 4.8 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+45.7points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+45.3points ahead
- Real work tasks from 44 professionsGDPVal+29.8points ahead
Where Step 3.7 Flash pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+5.1points ahead
Every test, side by side
All 21 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.8
- Terminal-Bench 2.1Claude Opus 4.8 by 45.384.639.3+45.3
- Terminal-Bench HardClaude Opus 4.8 by 22.758.335.6+22.7
- SWE-bench ProClaude Opus 4.8 by 12.969.256.3+12.9
- SciCodeClaude Opus 4.8 by 10.554.443.9+10.5
AgentsClaude Opus 4.8
- GDPValClaude Opus 4.8 by 29.847.818+29.8
- AA ApexAgentsClaude Opus 4.8 by 24.639.414.8+24.6
- τ-Bench V3 · BankingClaude Opus 4.8 by 22.234.212+22.2
ReasoningClaude Opus 4.8
- Humanity's Last ExamClaude Opus 4.8 by 27.348.721.4+27.3
- CritPtClaude Opus 4.8 by 18.620.92.3+18.6
- GPQA DiamondClaude Opus 4.8 by 11.19280.9+11.1
FactsClaude Opus 4.8
- AA-Omniscience · Non-hallucinationClaude Opus 4.8 by 45.760.715+45.7
- AA-Omniscience · AccuracyClaude Opus 4.8 by 2348.825.8+23
MathClaude Opus 4.8
- HMMT Feb. 2026Claude Opus 4.8 by 7.695.587.9+7.6
- AIME 2026Claude Opus 4.8 by 510095+5
Following instructionsStep 3.7 Flash
- IFBenchStep 3.7 Flash by 5.162.267.3+5.1
Other results5 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- AA-OmniscienceClaude Opus 4.8 by 66.128.8-37.3+66.1
- Artificial Analysis Coding IndexClaude Opus 4.8 by 34.774.339.6+34.7
- AA IntelligenceClaude Opus 4.8 by 22.341.819.5+22.3
- AA Agentic IndexClaude Opus 4.8 by 20.942.621.7+20.9
- τ²-Bench Telecom (AA run)Step 3.7 Flash by 4.194.498.5+4.1
Questions people ask
Which is better, Claude Opus 4.8 or Step 3.7 Flash?
Claude Opus 4.8 wins six of the seven areas where both have results: coding, agents, reasoning, facts, math and long documents. Step 3.7 Flash wins following instructions, and costs 96% less.
Which is better for coding?
Claude Opus 4.8. It wins 4 of the 4 coding tests both models report; Step 3.7 Flash wins none.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Step 3.7 Flash costs $0.20 and $1.15. That makes Step 3.7 Flash about 96% cheaper for the same work.
How do you compare the two?
We use the 21 benchmark tests both models have published scores on. The verdict counts the 16 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 5 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.